Profiling the Organic Emissions from a Light-Duty Direct ...Fundamentals of Combustion. Organised by...
280
Profiling the Organic Emissions from a Light-Duty Direct Injection Diesel Engine over a range of Speeds and Loads by Anthony Richard Collier A thesis subm itted to the University of Plymouth in partial fulfilment for the degree of Doctor of Philosophy Department of Environmental Sciences Faculty of Science In collaboration with Perkins Technology, Peterborough March 1995
Transcript of Profiling the Organic Emissions from a Light-Duty Direct ...Fundamentals of Combustion. Organised by...
Profiling the Organic Emissions from a Light-Duty Direct Injection Diesel Engine over a range of Speeds
and Loads
by
Anthony Richard Collier
A thesis submitted to the University of Plymouth in partial fulfilment for the degree of
Doctor of Philosophy
Department of Environmental Sciences Faculty of Science
In collaboration with Perkins Technology, Peterborough
March 1995
; <,t: rLYI\t..:.JUTH
Hem.No. 2..
Date · 2 1 AUG 1995
Ill
I REFERENCE ONLY I
Profilim:; the Organic Emissions from a Lig)lt-Duty Direct Injection Diesel Engine over a range of Speeds and Loads
Anthony Richard Collier
Abstract
Diesel engines account for a large percentage of the particulates in UJban city environments. Polycyclic aromatic compounds (PAC), some proven carcinogens, have been found on diesel particulates. The trace level nitro-PAC emissions, such as 1-nitropyrene and dinitropyrenes, contribute a large proportion of the mutagenicity in the particulates; in the case of 1-nitropyrene between 10 to 40% of the total mutagenicity of the particulate has been claimed. The potential health hazards of PAC require the levels and sources of such emissions to be evaluated over a range of speeds and loads.
PAC emissions are dependant on the engine specification, such as normally aspirated compared with turbocharged, and the operating conditions (speed and load). The effect of such variables can be determined using emission profiling, in which profiles of the exhaust are compared at varying engine powers. In this way the effect of speed and load on the combustion efficiency can be established.
Identification of PAC sources may be further complicated when engine sampling systems, such as the conventional dilution tunnel/filter system, are prone to artefact formation. This is especially relevant to secondary nitro-PAC emissions, which are prone to forming as artefacts of the filter installed in the dilution tunnel. In this study, organic emission maps were constructed using the unique total exhaust solvent-stripping apparatus (fESSA) developed at the University of Plymouth. TESSA allowed rapid sampling with a minimum potential for artefact formation. The close proximity of TESSA to the engine allowed the role of the combustion chamber in the formation of emissions to be evaluated.
Primary organic emissions, such as pyrene, are derived from survival of compounds in the fuel/oil and by combustion generation. Establishment of emission maps for the primary emissions are vital to resolving the formation of secondary emissions, such as 1-nitropyrene. Profiling of primary emissions sampled 26 different spe<lds and loads using TESSA (sampling times as low as IS seconds). Following simple work-up, quantification of the o-alkanes was by gas chromatography with flame ionisation detection and PAH by gas chromatography/mass spectrometry operated in electron impact mode. Emissions were expressed as a recovery of the compound emitted as a percentage of the same compound entering the chamber in the fuel. The o-alkane and PAH emission maps correlated with the gaseous unbumt hydrocarbon emissions, indicating that fuel survival was an important source of emissions, whereas lubricating oil contnbutions were minimal. Fuel survival contributions decreased with load; at 1000 rpm the average PAH survival of 0.95% at idling decreased to 0.2% at fuU load. High survivals under idling was a consequence of the low chamber temperatures and air:fuel ratios mixed beyond the lean flammability limits, whereas at full load, the high temperatures resulted in the greatest combustion efficiency. The o-alkane emission trends replicated those of PAH; atiOOO rpm the average o-alkane survival was 0.48% at idle compared with 0.084% at full load. Correlations between the distnbution of the emissions and fuel at high load, suggested fuel survival unchanged was responsible. At low loads the exhaust/fuel PAH ratios were more varied, with the range of percentage recoveries at low loads increasing with speed (difference between percentage recovery of fluorene and phenanthrene at idling for 1000 rpm and 3000 rpm was 0.05 and 0.18 respectively).
At high load, the combustion environment can be envisaged as producing areas under which complete combustion and survival unchanged occur. In between the complete combustion and unbumt fuel zones, a narrow range of temperatures and time for combustion reduce the opportunity for combustion generation reactions and/or preferential survival. Low load at 3000 rpm may increase the intermediate zone, allowing preferential survival and/or combustion generation reactions to evolve. Possible pyrolytic cracking of o-alkanes and demethylation of P AH at low loads and 3000 rpm was evident. The optimum time, swirl, and temperature for efficient combustion at low loads was generated at 2000 rpm to 2500 rpm, whereas at the higher temperatures corresponding to high loads, the effect of speed was much smaller. The primary emissions map show engineering improvements, particularly at low loads, could be implemented to lower the PAH emissions. The correlation between the emissions and fuel input suggest modifications to the PAH content of fuels may lower emissions.
The formation of secondary nitro- and oxy-PAC emissions is by transformation of primary emissions. In the case of nitro-PAC, nitration has been proposed to occur by free radical processes between PAH and oxides of nitrogen, NO,, within the chsmber and by electrophilic substitulion of PAH surviving combustion by nitrogen dioxide and nitric acid, via the nitronium ion. The combustion contnbution to nitro-PAC emissions was investigated using an upgraded TESSA system, and 3 speeds at low, mid, and high NO, for each speed were sampled. FoUowing initial extraction, concentration and clean-up of the samples, the nitro-PAC fractions of interest were isolated by nonnal phase high performance liquid chromatography. The nitro-PAC were identified and quantified by gas chromatography with electron capture detection, and gas chromatography/mass spectrometry operated in the negative ion chemical ionisation mode (the detection limits for both analytical systems was of the order of 40-50 pg of nitro-PAC standards injected). The profiling indicated that a proportion of the fuel underwent nitration within the combustion chamber across a range of speeds and loads. The extract concentrations (average of 5.3 ppm) found in this study were much lower than those previous found (ranging from SS to 2280 ppm). The majority of the previous studies relied on sampling using dilution tunnel/filter systems, for which post combustion contributions are simulated; suggesting that a major source of nitro-PAC is derived from post-a~mbustion nitration of PAH surviving combustion, some of which may be artefacts of the filter. Different speeds produced different trends for nitro-PAC emissions with respect to engine load_ It was not until the high temperature speed of 3500 rpm was reached that both NO, and nitro-PAC increased with load (R-sq= 0.989 & p=0.067 for 1-nitronaphthalene & NOJ.
Nitro-PAC emissions at 3500 rpm were primarily the result of combustion chamber nitration of PAH at high NO,. In the case of 1-nitropyrene, there was strong evidence to support pyrosynthetic contributions to the pyrene mass, which in turn became nitrated. The nitro-PAC emissions at low loads were the result of post-combustion nitration of PAH surviving combustion with the nitronium ion. The correlation of the PAH precursors to nitro-PAC in the fuel and nitro-PAC emissions suggest fuel modifications may to some extent lower the nitro-PAC ernissions. The combustion generation of nitro-PAC at high engine powers may require post-combustion after treatment.
Acknowledgements
I would like to thank my supervisors, Dr Colin Trier, Dr Mike Rhead, and Dr Murray Bell for their guidance and support throughout the research. I also thank the other members of the diesel research group at the University, in particular Professor David Fussey, Dr Paul Tancell and Robin Pemberton.
I wish to thank the staff at Perkins Technology for their advice in interpreting the results from an engineering viewpoint. Special thanks goes to Fred Brear for his assistance whilst at Perkins.
I must thank all the support from the staff at the University of Plymouth. I would mention in particular Roger Shrodzinski for his help with the GC/MS work and Dr Anthony Gachanja for his help with the chemiluminescence experimentation. I also thank Roy Cox and Don Ryder in the Thermodynamics laboratory for helping with maintaining and upgrading the engine sampling facilities.
I also wish to thank the staff, particularly Jim Carter, at the GC/MS facility at the University of Bristol for help and guidance with the use of their GC/MS facilities.
Finally I express many thanks to my family for their support throughout the research. A special thanks goes to Rachel Rossitter for helping me through the research.
Author's Declaration
At no time during registration for the degree of Doctor of Philosophy has the author been registered for any other University award.
Throughout the research I regularly attended and presented research seminars at the University of Plymouth. I also attended a number of key external conferences:
Emissions. Organised by UniCEG. Shell Thornton Research Centre, 22/9/93.
Petroleum Analysis. Organised by The Chromatographic Society. Shell Thornton Research Centre, 10/11/93.
Alternative Fuels. Organised by UniCEG. Coventry University, 20/12/93.
Fundamentals of Combustion. Organised by UniCEG. University of Warwick, 13/4/94
Urban Air Pollution and Public health. Organised by the Environmental Change Research Centre, University College London. 23/9/94.
Lecture given at the Fundamentals of Combustion Conference: Relationship between Fuel and Emissions in Diesel Engine Combustion.
Poster presented at the Urban Air Pollution and Public Health conference: Nitro-Polycyclic Aromatic Compound Profiles for a Light-Duty Diesel Engine.
I have also frequently consulted with external institutions and the following papers were prepared for publication:
Collier, A.R. (1994) Organic emissions from light-duty diesel engines. Proceedings of a Seminar: Urban Air Pollution and Public Health. Environmental Change Research Centre. University College London,
Collier, A.R., Rhead, M.M., Trier, C.J., & Bell, M.A. (1995) Polycyclic Aromatic Compound Profiles from a Light-Duty Direct Injection Diesel Engine. Fuel. 74 (3), 362-367.
Signed~V. Date . i¥./."i;ii'f._.
Contents Chapter 1 - Introduction
1.1 The Emergence of High-Speed Passenger Car Diesels
1.2 Environmental Impact of Diesel Emissions
1.3 Legislation and Control of Diesel Emissions
1.4 Diesel Combustion and the Formation of Organic Emissions
1.5 Profiling of lhe Organic Emissions from Diesels
Chapter 2 - Review of Diesel Emission Profiling
2.1 Review of Profiling Techniques
2.1.1 Sampling of Engines for Organic Emissions
2.1.2 Labomtory Work-up & Analytical Techniques for Organic Compounds
2. 1.2.1 Extraction and Concentmtion of Organics
2.1.2.2 Fmctionation of Organic
2.1.2.3 Detection of Organics
2.2 Review of Previous Profiling Studies
2.1.1 PAH Profiling
2.1.2 Alkane Profiling
2. 1.3 Nitro-PAC Profiling
2.1.4 Oxy-PAC Profiling
2.3 Summary of Emission Profiling & Aims of Investigation
Chapter 3 - Experimental Procedures & Facilities
3.1 Facilities and Fuel/Oil Description .
3. 1.1 Engine Details
3.1.2 Engine Management
3.1.3 Total Exhaust Solvent-Stripping Apparatus (TESSA)
3.1.4 Fuels and Oils
3.2 Profiling of lhe Primary Emissions using the initial TESSA configumtion
3.2.1 Engine Sampling for Primary Emissions
3.2.2 Sample work-up
3.2.2.1 Internal Standards for Aromatics
3.2.2.2 Extraction and Concentmtion of TES
3.2.2.3 Open column Silica Gel Fmctionation
3.2.2.4 Fuel work-up
3.2.2.5 Lubricating Oil work-up
3.2.2.6 Recovery of Aromatic Internal Standards
3.2.2. 1 Labomtory losses for Selected n-Aikanes
3.2.3 Quantification of n-Aikanes by Gas Chromatogmphy with Flame Ionisation Detection
. I
6
10
13
11
19
19
19
23
23
24
24
28
28
31
39
48
50
53
53
53
53
53
56
51
51
60
60
62
62
63
63
64
65
66
3.2.4 Quantification of Major PAC by Gas Chromatography/Mass Spectrometry operated in
Electron Impact Mode
3.3 Relocation and Modification of TESSA
3.3.1 Improvements to the Engine Sampling Rig
3.3.2 Initial testing of the upgraded TESSA
3.4 Profiling of the Secondary Emissions using the upgraded TESSA
3.4.1 Engine Sampling for Nitro-PAC Profiling
3.4.1.1 Engine Sampling for Nitro-PAC at 2500 rpm
3.4.1.2 Engine Sampling for Nitro-PAC at 1500 rpm & 3500 rpm
3.4.2 Sample Work-up for Nitro-PAC Profiling
3.4.2.1 Extraction and Concentration of TES for Nitro-PAC Profiling
3.4.2.2 Clean-up of TES
3.4.2.2.1 Octadecylsilane Cartridge Clean-up
3.4.2.2.2 Silica gel Clean-up
3.4.2.3 High-Performance Liquid Chromatography Fractionation
3.4.2.3.1 Initial HPLC Fractionations and Method Development
3.4.2.3.2 HPLC Fractionations of TES collected at 1500 rpm & 3500 rpm
3.4.2.4 Fuel and Oil work-up
3.4.2.5 Recovery of Nitro-PAC for Laboratory Work-up
3.4.3 Analysis Nitro- and Oxy-PAC
3.4.3.1 Evaluation of Detection Systems for Nitro-PAC Analysis
3.4.3.1.1 Gas Chromatography with Nitrogen-Phosphorous Detection
3.4.3.1.2 Reverse Phase HPLC with Fluorescence/Chemiluminescence Detection
3.4.3.1.3 Gas Chromatography/Mass Spectrometry operated in the Electron Impact Mode
3.4.3.1.4 Gas Chromatography with Electron Capture Detection
3.4.3.1.4.1 Analysis of Mononitro-PAC HPLC Fractions by GC-ECD
3.4.3.1.4.2 Analysis of Dinitro-PAC HPLC Fractions by GC-ECD
3.4.3.1.4.3 Analysis of Nitro-oxy-PAC HPLC Fractions by GC-ECD
3.4.3.1.5 Gas Chromatography/Mass Spectrometry operated in the Negative Ion Chemical
Ionisation mode
69
75
75
79
80
80
80
83
85
85
86
86
89
90
91
93
93
95
95
96
96
97
101
101
103
108
108
Ill
3.4.3.2 Gas Chromatography/Mass Spectrometry operated in Electron Impact mode for Oxy-PAC 115
3.4.3.3 Analysis ofn-Aikanes by Gas Chromatography with Flame Ionisation Detection 115
3.4.3.4 Analysis of major PAC by Gas Chromatography/Mass Spectrometry in the Electron
Impact Mode .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
.. 11
Chapter 4- Primary Organic Emissions
4.1 The Contributions of Fuel and Oil Survival to the Emissions
4.2 Profiling of the Primary Emissions using the initial TESSA configuration
4.2.1 The Concept of Percentage Recoveries
4.2.2 The Effect of Engine Load on the Percentage Recovery of Primary Organics
4.2.3 The Effect of Engine Speed on the Percentage Recovery of Primary Organics
4.2.4 Combined Effect of Engine Speed and Load on Primary Emissions
4.2.5 Comparison of the n-Aikanes with the Major PAC emissions
4.2.6 The contribution of combustion generation to the organic emissions
4.3 Comparison of Primary Emissions collected using Initial & Upgraded TESSA
4.3.1 Combustion of Fluorene at low loads and low speed
4.3.2 Effect of Engine Load on the Percentage Recovery of Primary Emissions
4.3.3 Effect of Engine Speed for the Combined Sampling Sets
4.4 Summary of Major Organic Emissions
Chapter 5 -Secondary Nitro- & Oxy-PAC Emissions
5 .I The Contnbution of Different PAC classes to the Emissions
5.2 Nitro-PAC Emissions
5.2.1 Analysis of the Mononitro-PAC HPLC fractions by Gas Chromatography with Electron
Capture Detection
5.2.1.2 Analysis of the Mononitro-PAC HPLC fraction derived at High Speed and High Load by Gas
118
118
128
129
134
142
145
149
!52
160
160
!62
!65
167
169
169
171
172
Chromatography/Mass Spectrometry operated in Negative Ion Chemical Ionisation Mode 178
5.2.1.3 Comparison of the Levels and distribution ofNitro-PAC found using TESSA with Previous
Studies
5.2.1.4 The Effect of Engine Speed and Load on the Mononitro-PAC Emissions
5.2.1.5 Factors controlling the Nitration in the Combustion Chamber
5.2.1.6 Nitration mechanisms involved
5.2.2 Dinitro-PAC & Nitro-oxy-PAC Profiles for the Prima Diesel Engine
5.3 Oxy-PAC Emissions
5.4 Sununary of Findings for Secondary Nitro- and Oxy-PAC Emissions
Chapter 6 -Final Discussions, Conclusions, & Future Work
6.1 Final Discussions
6.1.1 Primary Organic Emissions
6.1.2 Secondary Organic Emissions
6.2 Conclusions from research
6.3 Future Research Proposals
Literature Cited
Appendix- Nitro-PAC Calculations
... Ill
!83
189
195
198
207
207
213
215
215
216
223
232
234
236
AI
List of Figures
Chapter 1 - Introc1uctjoo 1.1 The superior fuel economy of diesel engines compared with petrol engines, especially for
urban driving. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 The configuration of direct injection (DJ) and indirect injection (ID I) combustion chambers. 3 1.3 The Perkins Prima high speed direct injection (HSDI) engine. S 1.4 Structures of polycyclic aromatic compounds (P A C) common to diesel emissions 8 1.5 The trade-off between emissions of particulate and oxides of nitrogen (NOJ 12 1.6 The four cycles of the diesel engine, consisting of the induction (a), compression (b),
power (c), and exhaust (d) stages .............................................................................. 14 I. 7 Axisymmetric idealization of a diesel fuel spray. The lower section of the figure shows the lines
of constant equivalence ratio (stoichiometric air:fuel ratio I actual air: fuel ratio, denoted by cj!). The areas of the fuel spray corresponding to the rich and lean flammability limits are indicated.. 15
Chapter 2 - Review of Diesel Emission Profiling 2.1 A typical dilution tunnel and filter collection engine sampling system 20 2.2 The total exhaust solvent-stripping apparatus (TESSA). 22 2.3 An idealised example of the chromatography for three model compounds (a) and the resulting
chromatogram (b)................................................................................................ 25 2.4 The effect of speed and load on the PAH emissions from ID! (a) and Dl (b) diesels. 30 2.5 Effect of engine load on the PAH and Alkyl-PAH emissions from a light-duty Dl diesel. 32 2.6 Effect of engine speed on the survival of PAH from a medium-duty DI diesel. 34 2.7 The ratio between PAC in TES relative to the PAC in the fuel for a) low engine power and
b) high power..................................................................................................... 36 2.8 The effect of engine load on the survival of u-Aikanes from a light-duty ID! diesel. 38 2.9 The effect of engine conditions on 1-nitropyrene emissions from a heavy-duty diesel. 41 2.10 Effect of different segments of a FTP cycle on emissions of 1-nitropyrene (a), emissions of
NO, (b), and temperature (c).............................................................................. 42 2.11 Effect of engine load on the nitro-PAC emissions from a heavy-duty turbocharged diesel 45 2.12 The effect of engine load and Speed on the 1-nitropyrene emissions from a medium-duty Dl
diesel............................................................................................................ 47 2.13 Effect of engine load on the Oxy-PAC emissions from a heavy-duty diesel. 49
Chapter 3 - Experimental Procedures & Facilities 3.1 The operation of the solvent delivery system (a) and exhaust transfer mechanism (b) used for
engine sampling with TESSA.. ... .. ....... .. ...... .. . . . . . . . ... ... . .. .. . . ... . . .. .. . .. . . . . . . . . . . . . .. . . . . . .. . .. . . . . 55 3.2 The overall experimental procedure for the profiling of the primary organics. 58 3.3 Gas chromatogram of the u-alkane standard mix. GC conditions: DB-5 column, SO"C for I min.,
IO"C/min. up to 270'C............................................................................................ 67 3.4 Calibration curve for beptadecane using gas chromatography with flame ionisation detection. 68 3.5 The identification of PAC on a retention and molecular ion basis. GC conditions: PTE-5 column,
40"C for I min., 6"Cimin. up to 300'C, held for IS mins...... ... . . .. . . .. .. .. . .. . . . . . . . . . . . .. . . . .. . . . .. . . 70 3.6 The confirmation of pyrene by matching the spectra in the sample (a) to that of the standard
spectra stored int he computer library (b)................................................................. 71 3. 7 Identification of phenanthrene and methylphenanthrenes by the respective molecular ions, 178 =
phenanthrene, 192 = monomethylphenanthrenes, 206 = dimethylphenanthrenes, and 220 = trimethylphenanthrenes. GC conditions: PTE-5 column, 40"C for I min., 6"C/min. up to 300"C, held for 15 mins.................................................................................................... 72
3.8 Total ion chromatogram of the standard PAC used for the primary PAC profiling. GC conditions: PTE-5 column, 4o•c for I min., 6"C/min. up to 300'C, held for 15 mins............................. 74
3.9 The overall experimental procedure aimed at profiling the secondary nitro- and oxy-PAC. 81 3.10 The effect of engine speed and load on the gaseous oxides of nitrogen, NO,. 82 3.11 The evaluation of the acidic aqueous methanol phase for the formation of nitronaphthalenes.
The comparison of the nitration check (a) to both the solvent blank (b) and the mononitro-PAC standard mix (c) proved there was no such nitration occurring. (Chromatographic conditions given in Section 3.4.3.1.4.1.)........................................................................ 87
. IV
3.12 The evaluation of the acidic aqueous methanol phase for the formation of 3-nitrofluoranthene and 1-nitropyrene. The comparison of the nitration check (a) to both the solvent blank (b) and the mononitro- PAC standard mix (c) proved there was no such nitration occurring. (Chromatographic conditions given in Section 3.4.3.1.4.1.)...... .. ... .... ... ... .. . . . . . .. ... .. . . . .. . .. 88
3.13 The chromatogram of the HPLC fractionation standard mix and the establishment of the initial HPLC fraction elution windows. HPLC conditions: Spherisorb silica column, I 00% hexane
(5 min.), increased to 100% DCM (20 min.), held for 10 min., increased to 100% acetonitrile (10 min.), and held for 10 min. Flow rate of 4 ml/min.................................................. 92
3.14 The HPLC fractionation of the aromatic fraction from the ODS clean-up of TES collected at 2500 rpm and low load. HPLC conditions: Spherisorb silica column, 100% hexane (S min.), increased to
100% DCM (20 min.), held for 10 min., increased to 100% acetonitrile (10 min.), and held for 10 min. Flow rate of 4 ml/min... ........ ... ......... .... ... ....... .. ... ..... .. . .. .. . .. . . . .. . . . . . . . .. . . 94
3.15 The sensitivity of fluorescence detection with on-line pre-column reduction for the determination of 1-nitropyrene (2.5 pg injected) in the reduced form of 1-aminopyrene. HPLC conditions: Spherisorb ODS-2 column, isocratic elution with acetonitrile: water (80:20), and ammonium acetate (30 mM). Ex1 = 260 nm and Em, = 430 nm. Flow rate of I ml/min... .. . . . . .. .. . .. ... .. .. ... .. . .. .. 98
3.16 The sensitivity of the bis (2,4-<linitrophenyl)oxalate (DNPO) chemiluminescence reaction to the detection of 1-aminopyrene (20 ng injected each time) using flow-injection analysis. Flow rate of I ml/min for mobile phase (lOO% acetonitrile) and O.S ml/min for each reagent stream (DNPO and H20,)......................................................... ................................. 100
3.17 The low sensitivity of gas chromatography/mass spectrometry operated in the electron impact mode to mononitro-PAC standard mix. GC conditions: Rt,-S column, 60"C for I min., S"C/rnin. to 300"C, held for 10 min.............................................................................. 102
3.18 The high sensitivity of gas chromatography with electron capture detection to a) the nitronaphtbalenes, and b) 3-nitrofluoranthene and 1-nitropyrene. GC conditions: Supelchem PTE-5 column, 60"C for I min., S"C/min to 300"C, held for 10 min........................... 104
3.19 The establishment of the integration temperature windows for nitro-PAC using the mononitro-PAC standard mix. GC conditions: Supelchem PTE-S column, 60"C for I min.,
S"C/min. to 300"C, held for 10 min................................................................... lOS 3.20 The confirmation of the nitro011phthalenes in the mononitro-PAC HPLC fraction collected
at high load and 1500 rpm (a) by comparing the nitronaphthalene temperature window (b) with that of the same san1ple eo-injected with the mononitro-PAC standard mix (c)
(Chromatographic conditions given in Section 3.4.3.1.4.1)................................... .......... 107 3.21 Calibration curves for 1-nitronaphthalene and 1-nitropyrene using gas chromatography with
electron capture detection....................................................................................... I 09 3.22 The gas chromatogram of the dinitro-PAC standard mix. GC conditions: Supelchem PTE-5,
60"C for I min., IO"C/min to 180"C, S"C/min to 300"C, held for 10 min.................... 110 3.23 The gas chromatogram of the nitro-oxy-PAC standard mix. GC conditions: Supelchem PTE-5,
60"C for I min., IO"C/min to 200"C, S"C/min to 300"C, held for 10 min........................ 112 3.24 The sensitivity of gas chromatography/mass spectrometry operated in the negative ion chemical
ionisation and selective ion modes to a) 1- & 2-nitronaphthalene, b) 2-methyl-1-nitronaphtbalene, and c) 3-nitrofluoranthene and 1-nitropyrene. GC conditions: Supelchem PTE-5 column, 60"C for 1 min., S"C/min. to 300"C, held for 10 min......................................................... 113
3.25 The analysis of the mononitro-PAC HPLC fraction obtained at 3500 rpm and high load by gas chromatography/mass spectrometry operated in negative ion chemical ionisation mode, using selective ion monitoring (a) and the confirmation of 1-nitropyrene by scanning for the 247 molecular ion (b). GC conditions: Supelchem PTE-5 column, 60"C for I min., S"C/min. to JOO"C, held for 10 min............................................................................................. 114
3.26 Total ion chromatogram of the oxy-PAC standard mix. GC conditions: Rt,-5 column, 60"C for I min., S"C/min to 300"C, held for 10 min............................................................ 116
Chapter 4 - Primacy Organic Emissions 4.1 Comparison of the alipbatic fractions of diesel fuel (a), and lubricating oil (b), with the TESSA
sample (TES) at 1% of full load and 1000 rpm (c) (Chromatographic conditions given in Section 3.2.3).......................................................................................... 119
4.2 Close examination of the oil hump temperature window (a) to the same temperature range of the TESSA extracted sample (TES) at I% of full load and I 000 rpm (b) (Chromatographic conditions given in Section 3.2.3)............................................................................... 121
4.3 Exanlination of the alipbatic fractions of TESSA extracted samples (TES) for evidence of oil contributions at three power settings; SO% of full load & 1000 rpm (a), full load and 1000 rpm (b), and 1% of full load and 3000 rpm (c) (Chromatographic conditions given in Section.3.2.3)... 122
V
4.4 Comparison of the aromatic fractions of fuel (a), and lubricating oil (b), with the TESSA extracted sample (TES) at I% of fuU load and ISOO rpm (c) (Chromatographic conditions given in Section 3.2.4).................................................................................... 124
4.S Examination of the aliphatic fractions of sump oils for accumulation of unbumt fuel. Fresh oil (a), sump after SO hours use (b), and sump after !OS hours use (c) (Chromatographic conditions
given in Section 3.2.3)................................................................................... 125 4.6 Examination of the aromatic fractions of sump oils for accumulation of unbumt fuel. Fresh oil (a),
sump after SO hours use {b), and sump after lOS hours use (c) (Chromatographic conditions given in
Section 3.2.4)..................................................................................................... 126 4.7 Accumulation of phenanthrene and methyl derivatives in sump oil over time....................... 127 4.8 Interpretation of percentage recoveries for identification of the source of emissions.................. 131 4.9 Effect of engine load on the percentage recovery of selected PAC (a) and o-alkanes (b) at
1000 rpm.................................................................................................. 135 4.10 Effect of engine load on the emissions of unbumt hydrocarbons (UHC) (a), TESSA extracted
sample (TES) (b), and bosch smoke (c) at 1000 rpm................................................. 136 4.11 Effect of engine load on the percentage recovery of selected PAC (a) and o-alkanes (b) at
3000 rpm........................................................................................................... 143 4.12 Effect of engine load on the emissions ofunbumt hydrocarbons (UHC) (a), TESSA extracted
sample (TES) (b), and bosch smoke (c) at 3000 rpm...................................................... 144 4.13 Effect of engine speed and load on the percentage recovery of fluorene (a), dibenzothiophene (b),
phenanthrene (c), and pyrene (d).............................................................................. 146 4.14 Effect of engine speed and load on the percentage recovery of tetradecane (a), hexadecane (b),
octadecane (c), and eicosane (d)............................................................................. 148 4.1S Comparison of the percentage recovery of fluorene compared with hexadecane at 1000 rpm (a),
2000 rpm (b), and 3000 rpm (c).......................................................................... ISO 4.16 Comparison of the percentage recovery of phenanthrene compared with nonadecane at
1000 rpm (a), 2000 rpm (b), and 3000 rpm (c)........................................................ ISI 4.17 Effect of molecular weight on the percentage recovery of PAH (a) and o-alkanes (b)........... IS3 4.18 Comparison of dibenzothiophene with methyl-dibenzothiophenes at 1000 rpm (a) and 3000 rpm (b).
Comparison of fluorene with methyl-derivatives at 3000 rpm (c)...................................... IS4 4.19 Comparison of the o-alkane distribution in diesel fuel (a) with the TESSA extracted (TES) at I%
of full load & 3000 rpm (b)................................................................................ IS7 4.20 Effect of low loads on the percentage recovery of fluorene................................................ 161 4.21 Effect of engine load on the percentage recovery of selected PAC at ISOO rpm (a), 2SOO rpm (b),
and 3SOO rpm (c)............................................................................................. 163 4.22 Effect of engine load on the percentage recovery of selected o-alkanes at ISOO rpm (a)
and 3SOO rpm (b)......................................................................................... 164 4.23 Effect of speed on the percentage recovery of fluorene and hexadecane at low load...... 166
Chapter 5 -Secondary Nitro- & Oxy-PAC Emissions S.l Overall chemical composition of the aromatic fractions ofTES at ISOO rpm and a) low load,
b) mid load, and c) high load, determined using normal-phase high-performance liquid chromatography (Chromatographic conditions given in Section 3.4.2.3.2).... .... .... .. .. .. .. .. .... .. .. .. .. .. .. .. .. .... .. 170
S.2 The analysis of the mononitro-PAC HPLC fractions derived from sampling at ISOO rpm and across the load range, by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.1).............................................................. 173
S.3 The profile of the nitronaphthalenes at ISOO rpm and across the load range, determined using gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.1)............................................................................................. 174
S.4 The analysis of the mononitro-PAC HPLC fraction of diesel fuel by gllB chromatography with electron capture detection (a), and the verification of the lack of nitro-PAC in the fuel (b&c). Retention times for nitro-PAC standards: 1-nitronaphthalene (22.237 mins.), 2-methyl-1-nitronaphthalene (22.919 mins.), 2-nitronaphthalene (23.487 mins.), and 1-nitropyrene (42.0S4 mins.) (Chromatographic conditions given in Section 3.4.3.1.4.1)... 176
S.S The analysis of the mononitro-PAC HPLC fraction of sump oil (after SO hours use) by gas chromatography with electron capture detection (a), and the verification of the lack of nitro-PAC in the fuel (b&c). Retention times for nitro-PAC standards: 1-nitronaphthalene (22.237 mins.), 2-methyl-1-nitronaphthalene (22.919 mins.), 2-nitronaphthalene (23.487 mins.), and 1-nitropyrene (42.0S4 mins.) (Chromatographic conditions given in Section 3.4.3.1.4.1).............................................................................. .. ...... 177
. VI
5.6 Distribution of the nitronaphthalenes in TES collected at 3500 rpm and high load, determined by gas chromatography/mass spectrometry (GC/MS) operated in the negative ion chemical ionisation mode (a) compared with the distribution of naphthalenes in the fuel, as determined by
GC/MS in the electron impact mode (b) (Chromatographic conditions given in Sections 3.4.3.1.5 and 3.2.4 respectively.................................................................................... 180
5.7 Distribution of molecular ions 211 and 225 in the mononitro-PAC HPLC fraction collected at 3500 rpm and high load. Determined using gas chromatography/mass spectrometry operated in the negative ion chemical ionisation mode (Chromatographic conditions given in Section 3.4.3.1.5)...................................................................................................... 181
5.8 Distnbution of molecular ions 223 and 237 in the mononitro-PAC HPLC fraction collected at 3500 rpm and high load. Determined by gas chromatography/mass spectrometry operated in the negative ion chemical ionisation mode (Chromatographic conditions given in Section 3.4.3.1.5).. 182
5.9 The effect of engine load on nitro-PAC at 1500 rpm 190 5.10 The effect of engine load on nitro-PAC at 2500 rpm 191 5.11 The effect of engine load on nitro-PAC at 3500 rpm 192 5.12 The combined effect of speed 811d load on the emissions ofa) 1-nitronaphthalene and b) 1-
nitropyrene........................................................................................................ 193 5.13 The investigation of correlations between the nitro-PAC and NO, emissions at a) 1500 rpm
811d b) 3500 rpm................................................................................................ 197 5.14 Emissions of 1-nitronaphthalene (a), naphthalene (b), NO, (c) relative to the fuel burned at 1500
rpm................................................................................................................. 200 5.15 Emissions of 1-nitropyrene (a), pyrene (b), NO, (c), relative to the fuel burned at 1500 rpm 201 5.16 Emissions of 1-nitronaphthalene (a), naphthalene (b), NO, (c) relative to the fuel burned at 3500
rpm................................................................................................................ 203 5.17 Emissions of 1-nitropyrene (a), pyrene (b), NO, (c) relative to the fuel burned at 3500 rpm 204 5.18 Emissions of 1-nitropyrene (a), pyrene (b), NO, (c) relative to the fuel burned at low load 205 5.19 Emissions of 1-nitropyrene (a), pyrene (b), NO, (c) relative to the fuel burned at high load 206 5.20 The gas chromatographic 811alysis of the dinitro-PAC HPLC fractions collected at 3500
rpm and a) low load, b) mid load, and c) high load, by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.2)........................ 208
5.21 The gas chromatographic 811alysis of the nitro-oxy-PAC HPLC fractions collected at 3500 rpm B11d a) low load, b) mid load, 811d c) high load, by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.3)................... 209
5.22 The 811alysis of the oxy-PAC HPLC fractions of TES collected at low engine power (a) and high engine power (b), by gas chromatography/mass spectrometry operated in the electron impact mode (Chromatographic conditions given in Section 3.4.3.2)............................. 211
5. 23 The 811alysis of the polar fractions from the silica gel cle811-up of TES collected at low engine power (a) 811d high engine power (b), by gas chromatography/mass spectrometry operated in the electron impact mode (Chromatographic conditions given in Section 3.4.3.2)...................... 212
Chapter 6- Final Discussions Conclusions, & Future Work 6.1 Concept of different temperature and time zones necessary for complete combustion. High loads
result in the highest temperatures, resulting in a large complete combustion zone (ZJ, and a much smaller zone in which fuel sutvives intact (Z.). The gradient in tenns of the temperature 811d time existing in the intermediate combustion zone (ZJ is narrow. AI low load and high speed the unbumt fuel zone is greater and the fBI1ge of temperatures and time available for combustion is large, enabling preferential sutvival 8lldlor combustion chamber generation of PAC.. ....... ........................ 224
6.2 Two zone nitration concept, such that the free radical combustion generation of nitro-PAC is favoured at high loads and high speeds (a), whereas electrophilic substitution of sutviving PAC is favoured at low loads 811d low speeds (b)................................................................... 230
vii
Ljst of Tables
Chapter 1 - Introduction 1.1 Chemical compounds and species found io diesel emissiollB 7 1.2 Carcioogenicity associated with different PAC Classes 9 1.3 Diesel Engioe Passenger Car European Legislation 10
Chapter 2 - Review of Profiling of Organic Emissions from Light-Duty Diesels 2.1 Detection systems for o-Aikaoes and major PAH 26 2.2 Detection systems for nitro-PAC 27 2.3 PAH Emission Rates (mg/kg) io DI exhausts at two different air: fuel ratios and for four fuels
blends............................................................................................................... 33 2.4 Effect of two different engioe conditions on the Nilro-PAC Emissions from a heavy-duty diesel 43 2.5 Effect of Driviog ConditiollB on 1-Nilropyrene Emissions from Light- and Heavy-Duty diesels 46
Chapter 3 - Experimental Procedures & Facilities 3.1 Perkios Prima Engioe Details 54
3.2 Standard No. 2 Diesel Fuel Properties 56 3.3 Engioe sampliog series al 1000 rpm 59 3.4 Engine sampliog series at 2000 rpm 59 3.5 Engioe sampliog series at 3000 rpm 59 3.6 Results from Internal Standard & PAH Laboratory Recovery Experiment 61 3. 7 Internal standards added lo TES & Fuel for Primary Emission Profiliog 62 3.8 Laboratory Recovery for Internal Standards added to TES and Fuels 65 3.9 Laboratory Recovery of o-Aikaoes from Spiking Experiment 65 3.10 o-Aikaoe standard mix used for analysis and recovery experiments 66 3.11 Response factors of PAC lo Undecane 73 3.12 Engioe sampliog for Nilro-PAC at 2500 rpm 83 3.13 Engioe sampliog for Nilro-PAC at 1500 rpm 84 3.14 Engioe sampliog for Nilro-PAC at 3500 rpm 84 3.15 HPLC fraclionalion standard mix 91 3.16 Reproducibility of Retention Times (R,) for NP HPLC Fractionation 91 3.17 Laboratory recovery of nitro-PAC 95 3.18 Mononitro-PAC standard mix 96 3.19 Retention Index (RI) based on 1-nitropyrene 106 3.20 Response factors (R,) for PAC based on d10-Fluorene 117
Chapter 4 - Primacy Organic Emissions 4.1 Concentration of Selected o-alkanes in the A2 Fuel 120 4.2 Concentration of Selected PAC io the A2 Fuel 123 4.3 The Exhaust PAC emissions relative to the correspondiog fuel PAC consumed during sampling 132 4.4 The Exhaust o-alkane emissions relative to the corresponding fuel o-alkane consumed during
sampling............................................................................................................ 133 4.5 Previous studies on Fuel Survival for Diesel Engines 138 4.6 The effect of engioe load on the combustion and exhaust temperatures 139 4.7 Effect of Speed on the Percentage Recovery (PR) of PAC between 1% and 25% of Full Load 147 4.8 Effect of Speed on the Percentage Recovery (PR) ofn-alkanes between 1% and 25% of Full Load 147 4. 9 o-Alkane areas normalised to Hexadecane for a range of Speeds and Loads 159
Chapter 5 - Secondary Nitro- & Oxy-PAC Emissions 5.1 Nitro-PAC emitted from the Prima Engioe at different engioe conditions 178 5.2 Previous Nitro-PAC levels from Light-Duty Diesel Engioes 183 5.3 Previous Nitro-PAC levels from Medium- & Heavy-Duty Diesel Engioes 184
Chapter 6- Final Discussions, Conclusions & Future Work 6.1 Overview of the EmissiollB Profiling Research Programme 215
List of Plates Plate 1 The spray head allachmenl used to evenly disperse the solvents with the oncomiog exhaust 77 Plate 2 The transfer line from the engine lo TESSA 77 Plate 3 The overall engioe lest bay, showing the upgraded TESSA 78
viii
Chapter 1 - Introduction
1. I The Emergence of High Speed Passenger Car Diesels
The development of the diesel engine relies to a large extent on the greater thermal efficiency
associated with compression-ignition compared with spark-ignition (SI) used for petrol engines
(Amann et al. 1980, Haddad 1984, Kamimoto & Kobayashi 1991, Schindler 1992, and
Poulton 1994). The improved fuel economy of a diesel relative to a similarly powered SI
engine, in the region of 25% to 40% greater (Hillier 1991, Schindler 1992, and QUARG
1993}, results from the higher compression ratios and varying only the fuel supply to produce
the power (Hillier 1991 and Stone 1992). The mixing of fuel and air in the carburettor
generates large pumping losses at low loads for SI engines (Stone 1992). The difference in fuel
economy is greatest under urban driving conditions, where high air: fuel ratios are encountered
(Andrews 1986 and Poulton 1994) (Figure 1.1).
Commercial vehicle operators have long recognised the efficiency and reliability of the diesel
engine, resulting in diesel engines being used to power heavy goods vehicles (HGV), buses,
coaches, tractors, taxis, and light vans (Fukuda et al. 1992 and QUARG 1993). The diesel
engine used for commercial use was the open chamber or direct injection (PI) configuration
(Figure 1.2a). The fuel is injected directly into the chamber shortly before top dead centre
(TDC) whereupon the temperature of the compressed gases is sufficient to initiate combustion.
The necessary air motion, termed swirl, for efficient combustion is generated from the
chamber bowl and air inlet design.
Up until the mid-1980s only indirect injection (IDI) diesels were used in passenger cars
(Figure 1.2b). Fuel injected into the hot pre-chamber region is rapidly atomized and
combusted. The combustion generates high swirl and carries the remaining fuel charge into the
main chamber whereupon combustion is completed. Compared to the traditional heavy-duty
PI system, the IDis produced a smoother power delivery and were able to operate over a
wider speed range.
Early diesels were noisy, smelly, and much less responsive than SI engines, and also suffered
from cold start problems (Langley 1991 and Hillier 1991).
1
c 40 0 m 0> ..... Q) a.
30 (f)
~
:§_ >-E 20 0 c 0 tl Q)
Qi ::l
LL 10
Diesel
Petrol
Diesel/ Petrol
33.4
19.8
1.68
45.6
31.7
1.44
DDiesel
•Petrol
Figure 1.1 The superior fuel economy of diesel engines compared with petrol engines, especially for urban driving. (Andrews 1986)
2
Air inlet
Inlet & Outlet
. Fuel injector ~
Top oil ring liner
(a) DI
Combustion bowl
(b) IDI
-Fuel injector
Piston head recess
Figure 1.2 The configuration of direct injection (DI) and indirect injection (IDI) combustion chambers. (Monaghan 1981)
3
The power output is increased by upgrading the naturally air aspirated system to a
turbocharged inlet system, in which a greater density of air molecules are forced into the
chamber, thus enabling more fuel to be burned (Lilly 1984).
The IDI design suffers from large heat losses through the pre-chamber at high speeds, reducing
the fuel economy closer to that of SI counterparts (Monaghan 1981, Nahum 1989, and Hillier
1991). The problems associated with the IDis led to the development of high speed direct
injection (HSDI) diesel engines for passenger cars (Monaghan 1981, Nahum 1989, Hillier
1991, Belardini et al. 1993, and Horrocks 1993). One example of an advanced HSDI diesel
suitable for light-duty applications is the Prima engine developed by the Perkins Group, United
Kingdom (UK) (Figure 1.3a). The bowl-in-piston combustion cavity combined with the
helically shaped air inlet port enables high speed rotary swirl to be generated (Figure 1.3b),
resulting in high power and good fuel economy up to a rated speed of around 4500 rpm
(Hillier 1991).
The improved performance of the HSDI diesel engines combined with the fiscal benefits of
diesel motoring has led to increased market penetrations in most European countries. For
example, the market penetration of diesel cars increased from around 5% in 1988 to nearly
20% in the Autumn of 1993 within the UK market (QUARG 1993), and in France diesels have
reached 40% of the market share (Kerswill 1992). The increasing popularity of the diesel
engine in European countries is contrasted by almost no diesel sales in the United States (US)
(Henk et al. 1992). This is largely due to the very stringent particulate legislation implemented
in the US by the Environmental Protection Agency (EPA). The increasing proportion of
passenger cars utilizing diesel engines in the European markets makes an environmental
assessment of diesel emissions increasingly necessary.
4
Fuel ~ c:::==::-.. return
?l Fuel input to injector
(a) Perk:ins Prima DI combustion chamber
Swirl motion
(b) High-speed rotary swirl produced by inclined air inlet port
Figure 1.3 The Perkins Prima high speed direct injection (HSDI) engine. (Hillier 1991)
5
1.2 Environmental Impact of Diesel Emissions
The higher thermal efficiency of the diesel engine results in much less gaseous carbon
monoxide (CO) and hydrocarbons (HC) than equivalent powered SI engines (Williams et al.
1989, Stone 1992, and QUARG 1993). The higher fuel economy and corresponding lower
greenhouse gases per mile travelled may help to reduce global warming (Hammerle et al.
1991, Springer 1991, Schindler 1992, and QUARG 1993). However, the higher density of
diesel fuel and the possible warming effects of the particulates emitted from diesels may
narrow the difference between the contributions of diesel and SI engines to global warming
(QUARG 1993). The diesel engine can also run on a much wider selection of fuels compared
to SI engines (Shore 1986, Ziejewski et al. 1991, Zeijewski & Goettler 1992, Amann 1993,
and Shay 1993).
On the negative side, diesels emit more particulates than similar SI engines (Fukuda et al.
1992). The QUARG report (1993) on diesel emissions in the urban environment, estimated
that diesel particulates contribute as much as 87% of the total particulates in an urban
environment such as London. The particulates soil buildings, reduce visibility, whilst also
affecting atmospheric conditions (El-Shobokshy 1984, Trijonis 1984, Richards et al. 1986,
Ball 1987, Hamilton & Mansfield 1993, and QUARG 1993). The sub-micron size of the
particles allows deep penetration of the human respiratory systems (Cuddihy et al. 1984,
Scheepers & Bos 1992a, and Seaton et al. 1995). The health aspects of the particulates is of
great concern, since recent research suggests that the small size of the particles may in
themselves cause heart and lung problems, and exacerbate breathing disorders, such as asthma
(Wade & Newman 1993 and Seaton et al. 1995).
The high surface area of carbon particles facilitates adsorption and condensation of organic
compounds present in the gas phase, providing a medium for HCs to be transported deep into
the human body. An indication of the complexity of the organic fraction of diesel exhaust is
given in Table 1.1. The polycyclic aromatic compounds (PAC) have received a great deal of
attention since some PAC are established carcinogens in laboratory tests (!ARC 1989).
6
Table 1.1 Chemical compounds and species found in diesel emissions
Gas phase Particulate Phase
Acrolein, ammonia Heterocvclics and derivatives 1
Benzene, l, 3-butadiene Hydrocarbons (C,.-C .. ) and derivatives'
Fonnaldehvde, fonnic acid Polvcvclic aromatic comPOunds
Heterocvclics and derivatives'
1 Hvdrocarbons (C, -C,.) and derivatives
Hvdro11en cvanide, hvdro11en sulphide
Methane, methanol
Nitric acid, nitrous acid, oxides of nitrogen
Polvcvclic aromatic compounds
Sulnhur dioxide
Toluene
1 Derivatives include acids, alcohols, aldehydes, anhydrides, esters, ketones, nitriles, quinones, sulphonates and halogenated and nitrated compounds, and multi-functional derivatives.
(Adapted from IARC 1989)
The PAC are sub-divided into different chemical classes, such as polycyclic aromatic
hydrocarbons (PAH), alkyl-PAC, sulphur containing PAH (PASH), nitrogen containing PAH
(PANH), and nitro- and oxy-PAC (Figure 1.4). There has been a great deal of research into
whether the levels of PAC emitted from diesel engines are harmful, and these studies are
reviewed by Schenker 1980, Grasso et al. 1988, IARC 1989, Scheepers & Bos 1992a, Stober
1992, and Kingston 1994. Table 1.2 indicates the relative contributions of different sub-classes
of P AC and the relative carcinogenic potencies. The table illustrates the extent to which the
chemical structure influences the carcinogenicity, such that 6-methylchrysene is much less
potent than 5-methylchrysene (Melikian et al. 1983). Some members of the nitro-PAC group
are extremely hazardous in both mutagenic and carcinogenic testing (Beland et al. 1981,
Mermelstein et al. 1981, Pederson & Siak 1981, Salmeen et al. 1982, Andrews et al. 1983,
Rosenkranz & Mermelstein 1983, Wei & Shu 1983, Ohgaki et al. 1984, Tokiwa & Ohnishi
1986, and IARC 1989).
7
00 Naphthalene (P AH)
Dibenzothiophene (P ASH)
NOz
1-Nitropyrene (Nitro-PAC)
1-Methylphenanthrene (Alkyl-PAH)
Carbazole (P ANH)
0 I
Fluoren-9-one (Oxy-PAC)
Figure 1.4 Structures of polycyclic aromatic compounds (PAC) common to diesel emissions.
8
Table 1.2 Carcinogenicity associated with different PAC Classes
Compound Class Carcinogenicity in Carcinogenicity animals ~roup
Fluorene PAH I 3
Phenanthrene PAH I 3
Anthracene PAH I 3
Pvrene PAH I 3
Chrvsene PAH L 3
Benz( m )ovrene PAH s 2A
5-~ethvlchrvsene Alkyl-PAH s 28
6-~ethvlchrvsene Alkvl-PAH L 3
1-Nitronaohthalene nitro-PAC I 3
2-Nitrofluorene nitro-PAC s 28
1-Nitroovrene nitro-PAC s 28
6-Nitrochrvsene nitro-PAC s 28
1,3-Dinitropvrene nitro-PAC L 3
I ,6-Dinitroovrene nitro-PAC s 28
I, inadequate evidence; L, limited evidence; S, sufficient evidence; 2A, Group 2A - the agent is probably carcinogenic to humans; 28, Group 28 - the agent is possibly carcinogenic to humans; 3, Group 3 - the agent is not classifiable
to its carcinogenicity to humans (Adapted from IARC 1989)
Oxy-PAC and polar PAC are also possible candidates for the mutagenicity and carcinogenicity
associated with diesel extracts, although less information is available on which specific
compounds are responsible (Scheepers & Bos 1992a). Some oxy-PAC, such as fluoren-9-one,
are non-mutagenic, whereas others, such as benzo(a)pyrene-3,6-quinone and 9,10-
phenanthrenequinone are mutagenic in Ames testing (Scheepers & Bos 1992a).
The International Agency into Research on Cancer (IARC) classified diesel exhaust as a
probable carcinogen in 1989 (Group 2A), and in 1994 Pfeffer established a correlation between
diesel engine emissions and the carcinogenic risk in traffic-related sites in Germany. The PAC
content of diesel particulates may have been an important factor in the IARC carcinogenic
classification of diesel exhaust and the findings of Pfeffer.
9
The other emission problem associated with diesel engines is the production of higher
quantities of oxides of nitrogen, termed NO., compared to SI engines. This is particularly
relevant at the high load and speed working conditions for HGV transport travelling along
motorways. The QUARG report (1993) estimated that 43% of the NOx in London was
attributable to diesel engines. The NOx is emitted primarily as nitric oxide (NO), which is then
later oxidised to nitrogen dioxide (N02) in the environment (Kuhler et al. 1994 and Rudolf
1994). At high ambient levels N02 can cause health problems and NOx contributes to the
formation of photochemical oxidants, acid deposition and the global warming (QUARG 1993).
The NOx content of the combustion chamber and the exhaust may be of significance to the
formation of the hazardous nitro-PAC emissions (Section 1.4).
1.3 Legislation and Control of Diesel Emissions
European legislation has come a long way from 1963 when smoke emissions from diesel
vehicles were visually appraised by a panel of experts (Ball 1982). Vehicle legislation now
requires certification of new engine designs over a set of specified driving conditions or cycle
followed by derogation testing later in the vehicle's usage. The present European legislation
covers the emissions of HCs, CO, NO., and particulates (Table 1.3). The inclusion of a
division in Stage 11 between Dl and IDI engines arises from the HSDI engines producing as
much as three times more NOx and HCs, and two times more particulates than IDis (Belardini
et al. 1993).
Table 1.3 Diesel Engine Passenger Car European Legislation
11 • ._, . Year for approval CO !!lkm HC +NO /km Part'
911444/EEC 1991 2.72 0.97 0.14
Stage 2- IDI 199617 1.0 0.7 0.08 - DI 1.0 0.9 0.10
Sta2e 3 2000 0.5 0.5 0.04
(Adapted from Horrocks 1993)
10
The proposed tighter limits to control particulates and NO. are regarded as particularly difficult
to meet (Poulton 1994). This is due to measures, such as fuel injection retardation, aimed at
reducing the high temperature pre-mixed combustion phase associated with NO. production,
resulting in an increased diffusion burning phase within which particulates tend to form, and
vice versa (Signer & Steinke 1987 and Horrocks 1993) (Figure 1.5).
Exhaust gas recirculation in which a degree of the exhaust is mixed with the fresh air entering
the chamber, results in less oxygen available for combustion, and culminates in lower NO.
emissions (Amstutz & Del Re 1992). Improved combustion chambers can reduce both NO. and
particulates (Konno et al. 1992b & 1993), as can pilot two-stage and electronic fuel injection
(Signer & Steinke 1987 and Shakal & Martin 1990). Post-combustion after-treatment using
reductive catalysts can be used to lower NO. (Konno et al. 1992a and Davies et al. 1993),
whilst flow-through oxidation and particle traps can be used to reduce particulates (Signer &
Steinke 1987, Henk et al. 1992, and Davies et al. 1993). Such post-combustion techniques are
in the early stages of development with regards to light-duty passenger diesels (Andrews &
Gibbs 1993, Davies et al. 1993, and Zelenka & Herzog 1993). Increasingly computer
technology is being used to control the different emission control systems, allowing the
optimum conditions for the lowest emissions to be achieved at varying engine conditions
(French 1987, Carroll 1991, Bazari & French 1993, and Ladommatos et al. 1993).
Recently, legislators in the US and Sweden have imposed tight limits on the aromatic contents
of diesel fuels in an attempt to lower particulate emissions (Cooper et al. 1993 and Nikanjam
1993). The introduction of new fuel blends enables the benefits to be realised for all the
vehicles on the road, whilst engineering improvements take time to filter down into the car
market.
Presently, no legislation covers the emissions of P AC. However, increasing! y studies of the
urban environment are identifying the presence of PAC derived from diesel engines (Viras et
al. 1990, Aceeves & Grimalt 1993, Escriva et al. 1994, and Pfeffer 1994), providing pressure
on governments and legislators to evaluate the hazard associated with PAC.
11
1.0
5 10
N(\ (g/bhp-hour)
Figure 1.5 The trade-off between emissions of particulate and oxides of nitrogen (NOJ. (Kamimoto & Kobayashi 1991)
12
It follows that an understanding of the combustion and formation of organic emissions,
especially the hazardous nitro- and oxy-PAC, from diesel engines would enable control
strategies to be employed were PAH legislation to be introduced. As it is, the increasingly
tighter particulate emissions require the organic contribution to be assessed and if possible
reduced to lower the particulate mass collected as a whole.
1.4 Diesel Combustion and the Formation of Organic Emissions
The diesel engine works by using four cycles to draw air into the cylinder, compress the air
to generate high cylinder temperatures, inject and combust the fuel, and finally discharge the
chamber contents (Figure 1.6). The combustion phase itself can be separated into three stages;
the ignition delay, pre-mixed uncontrolled combustion, and controlled or diffusive periods
(Amann & Siegla 1982 and Kamimoto & Kobayashi 1991). During the ignition delay the fuel
is heated, vaporised and mixed with the swirling air. Chemical reactions occur, such as
pyrolysis and/or partial oxidation of the fuel. Combustion is initiated in the areas where the
air:fuel ratio relative to the ideal mix or stoichiometric ratio, termed the equivalence ratio (<I>),
is close to unity1 (Stone 1992). The ignition occurs in many places and there follows a rapid
pressure rise due to the uncontrolled combustion of the fuel (Schindler 1992). Following the
rapid uncontrolled combustion period, the remainder of the fuel must be injected, and in the
case of full load the remaining fuel constitutes a considerable amount of the total fuel injected.
During this diffusive phase of combustion a range of equivalence ratios, ranging from fuel rich
at the flame core to fuel lean at the spray edges, are generated (Figure 1. 7). Compounds will
only be completely combusted if they exist within certain equivalence limits, termed the
flammability limits. The upper flammability limit (<l>J refers fuel rich zones, whereas the
lower limit (<l>J equates to fuel lean areas (Stone 1992).
1Equivalence ratio (<I>) = stoichiometric air:fuel ratio I actual air: fuel
13
Key
In. = Air inlet Ex. = Exhaust outlet
(a)
'"enters cylind« through inlet pon
(c)
fin1ly 11omiJid full oil
Induction
..,.-- fuel delivered E.L 10 injiCIOI
-
high lit limP. lgnitH fuel 1nd liberiiH hill 10 procklce J)OWW llloU
(b)
lit being COfftCIIIII.O.. --+-~ temc-relu,. I ncl ptiiiU,.
lnCfiiM
(d)
Figure 1.6 The four cycles of the diesel engine, consisting of the induction (a) , compression (b) , power (c), and exhaust (d) stages. (HillU 1991)
14
'~Spray centerline
Too lean to burn where
•<•L /
lt.
Figure 1. 7 Axisymmetric idealization of a diesel fuel spray. The lower section of the figure shows the I ines of constant equivalence ratio (stoichiometric air: fuel ratio I actual air: fuel ratio , denoted by cf>). The areas of the fuel spray corresponding to the rich and lean fl ammability limits are indicated. (Ferguson 1986)
15
The organic fraction of the particulates are derived from organics in the gas-phase adsorbing
and condensing onto the particulates during the expansion and exhaust strokes. The origin of
the primary organics leaving the combustion chamber are derived from:
1) Survival of organics.
2) Combustion generation of organics.
The first source is derived from fuel and lubricating oil surviving combustion and entering the
exhaust. Standard diesel fuel is removed from the distillation process between 160°C - 390°C
and contains predominately aliphatic hydrocarbon chains of 9-28 carbon atoms (IARC 1989
and Scheepers & Bos 1992b). The fuel also contains a range of PAC, mostly 2-4 ring PAH
and their alkyl-derivatives (Williams et al. 1986 & 1989, Nelson 1989, Bundt et al. 1991, and
Trier et al. 1991). Fuel 'packets' mixed outside of the flammability limits may survive
combustion and enter the exhaust (Andrews et al. 1983 and Williarns et al. 1989). Similarly
lubricating oil leakages, containing scavenged organics, add to the particulates by surviving
combustion intact (Williams et al. 1987).
The second source is from combustion generation; either by pyrosynthetic reactions or by
pyrolysis of the fuel/oil components (Badger et al. 1964, Howard & Longwell 1983, Cole et
al. 1984, Williams et al. 1989, and Trier et al. 1990 & 1991). Pyrosynthetic reactions have
been linked with the formation of particulate, and P AC may be a by-product of the process
(Calcote 1981, Smith 1981, Serageldin 1981, Howard & Longwell1983, Williams et al. 1989,
and Burtscher 1992). The pyrosynthetic reactions take place inside the flammability limits
whereas pyrolysis occurs in fuel rich or low temperature areas (Williams et al. 1989).
Evidence for combustion generation can be masked by fuel/oil survival effects. Nelson (1989)
concluded that since the major aromatic species in diesel exhaust, phenylacetylene, styrene,
naphthalene, methylnaphthalenes, and acenaphthene, were not in the fuel they were a result
of combustion generation. More recently, carbon 14 e4C) experiments by both Pemberton
(1995) and Tancell (1995) have established the exact combustion generation contribution to
specific PAC. For example, Tancell et al. (1995a) determined that of the benz(a)pyrene
emitted from a light-duty DI diesel engine ea. 20%. was derived from combustion generation.
16
The emission of the primary organics can further react with surrounding oxygen and NO, to
form secondary emissions, such as nitro- and oxy-PAC. In the case of nitro-PAC, formation
is by nitration of PAH and PANH. Formation mechanisms for nitro-PAC have been proposed
by reactions with NO and N02 within the combustion chamber and by reactions with N02 in
the presence of nitric acid in the exhaust (Scheepers & Bos 1992b). Sump oil has been shown
to accumulate nitro-PAC, although evidence for the oil contributing to nitro-PAC emissions
has not been proved (Jensen et al. 1986). The excess air and high temperatures involved with
diesel combustion favour the oxidation of hydrocarbons, forming oxy-PAC such as fluoren-9-
one (Schuetzle et al. 1981 and Choudhury 1982).
Organic emissions can also originate from engine and/or exhaust deposits (Abbass et al. 1988),
and also from the engine sampling system itself. This is particularly relevant to secondary
emissions, such as nitro- and oxy-PAC, for which transformations of PAH can occur on the
filter used to collect the particulates (Section 2.1).
To reduce the contribution of both survival and combustion/exhaust generation to diesel
emissions may require different technical approaches. In order to lower emissions, an
understanding of the combustion of diesel fuel and subsequent organic emissions over a range
of speeds and loads is needed.
1 5 Profiling of Organic Emissions from Diesels
Diesel combustion at a particular speed and load results in a wide range of fuel:air ratios and
temperatures within the combustion chamber (Ferguson 1986). The fuel:air ratio ranges from
infinite at the injector nozzle to less than unity in the swirling air zone beyond the edge of the
flame front (Amann & Siegla 1982). Similarly, the areas of the combustion chamber associated
with the combustion ignition correspond to high peak gas temperatures, while areas further
from the flame front, such as the crevice volume space between the top oil ring liner and the
top of the chamber have much lower temperatures (Li 1982 and Signer & Steinke 1987). These
two parameters of fuel:air interaction and temperature will be central to the organic emissions
(Rubey et al. 1983 and Scheepers & Bos 1992b).
17
The time available for complete combustion will also be a factor in the combustion process
(Rubey et al. 1983), with less reaction time existing at high speed compared with lower speeds
(Zeijewski et al. 1991). The length and temperature of the exhaust section which the emissions
travel may also effect the final composition of the exhaust entering the environment (Williams
et al. 1985).
The combustion process will also be influenced by the type and configuration of diesel engine.
The type of fuel injector, injection timing, aspiration (normal, turbocharged, or supercharged),
and air inlet ports may all influence the combustion process (Scheepers & Bos 1992b and
Poulton 1994).
Emission profiling, in which emissions from an engine are characterised over varying speeds
and loads, enables the elucidation of relationships between the emissions and the outlined
combustion factors to be achieved for a specific engine set-up. The results can be compared
with other engines sampled in a similar way, and areas of significant emissions focused on.
Similarly, as compounds are classified as potentially harmful to human health, the role of
combustion and exhaust stages on such emissions can be followed with respect to speed and
load. Those PAC, especially those possessing nitro- and oxy- groups, which constitute a health
risk, require the levels of such compounds under different driving environments to be
established. This is particularly important for the low load and low speed driving conditions
associated with congested city centres.
It is of great importance that a detailed understanding of the formation pathways, including
precursor compounds (some of which may be not be carcinogenic but crucial to the subsequent
formation of compounds that are) which lead to high risk PAC are discovered. Information
concerning the degree to which the combustion chamber or the exhaust section determines the
PAC emissions under changing speeds and loads is vital for meaningful control strategies to
be formulated, in anticipation of possible future PAH legislation.
The different engine sampling and analytical techniques involved with emission profiling and
the previous profiling studies on diesel engines are reviewed in Chapter 2.
18
Chapter 2 - Review of Diesel Emission Profiling
This chapter reviews the research to date on the chemical characterisation and profiling of
organic emissions from diesel engines. The chapter is divided into three main sections. Firstly,
the sampling, work-up, and analysis techniques required for the profiling to proceed are
reviewed in Section 2.1. Previous studies on specific classes of organics are reviewed in
Section 2.2. The shortcomings highlighted in the review lead to outlining the aims and
objectives of this study in Section 2.3.
2.1 Review of Profiling Techniques
There have been a number of reviews on the sampling and analytical techniques available for
chemical characterisation of diesel emissions (Lee et al. 1981, Bjorseth 1983, Stenburg et al.
1983, White 1985, and Williams 1990). This review is aimed at highlighting those sampling
and detection techniques most appropriate to comprehensive emission profiling. The section
is divided into sampling (Section 2.1.1) and analytical techniques (Section 2.1.2).
2.1 1 Sampling of Engines for Organic Emissions
The standard method for sampling diesels, proposed by the US EPA, recommends sampling
the exhaust following dilution. The method was devised to simulate the dilution processes
following the exit of the exhaust into the environment (Schuetzle 1983). A typical dilution
tunnel configuration is shown in Figure 2.1. The exhaust is diluted by a known ratio of air
drawn through by the fan situated at the end of the tunnel. Dilution cools the exhaust and the
sample is collected on filters. The recommended collection of 52°C on the filters determines
the degree of dilution required for particular engines. The collection of gaseous phase
compounds in the exhaust which are not trapped by the filter, can be collected on a chemical
adsorbent trap behind the filter.
There are many forms of the dilution tunnel arrangements and each design can affect the
mixing in the tunnel and reactions in the transfer of the exhaust to the tunnel (Williams et al.
1985 and Trier 1988).
19
Filter""' l Dehumidifier:...--::;t:==t
Heater
l
Filter~
\Mixing Baffie
l
1 Gas Analysis
Exhaust from engine
l ---Vent
--Fan
Figure 2. 1 A typical dilution tunnel and filter collection engine sampling system. (Williams 1990)
20
An inherent problem associated with the system, is the possible reaction of organics held on
the filter with the incoming gaseous stream (Duleep 1980, Lindskog 1983, Hartung et al.
1984, Chan & Gibson 1985, and Trier 1988). Laboratory exposure of filters and reactor
vessels containing PAH to oxidising and nitrating agents have proved that transformations can
occur {lager & Hanu§ 1978, Pitts et al. 1978, and Hisamatsu 1986). Furthermore, the studies
by Herr et al. (1982) and Lach & Winckler (1988) showed that the exposure of pyrene on
filters to increasing higher exhaust N02 led to increasing 1-nitropyrene emissions.
The artefact potential for dilution tunnel/ftlter systems is exacerbated by the prolonged
sampling times needed for adequate sample collection. For example, a common transient test
cycle such as the US federal transient procedure (FTP} lasts for 23 minutes and may result in
a organic extract weight of ea. 50 mg (Bechtold et al. 1984). The lengthy sampling times not
only present Jogistical problems in terms of collecting a range of replicated samples on the
same day, but more importantly encourage artefact conversions of the sample on the filter.
This is especially relevant for the analysis of trace nitro-PAC, where repeated runs may need
combining in order to gain sufficient organic mass (Bechtold et al. 1984).
The other approach to diesel engine sampling is to directly sample the exhaust stream leaving
the combustion chamber. Early systems suffered greatly from ftlter degradation at the high
exhaust temperatures (Trier 1988). This led to cooling the exhaust by condensation methods
before collection on filters (Stenburg et al. 1983}, but the filter collection remained prone to
artefact problems.
To overcome the problems associated with dilution and filter systems Petch with eo-workers
(1987) devised a sampling system which dispensed filter collection, termed the total exhaust
solvent-stripping apparatus (TESSA). The system comprises a stainless steel tower into which
the exhaust is transferred at the base (Figure 2.2). The organic component of the exhaust is
rapidly extracted by the downward flow of solvent mix ((dichloromethane(DCM):methanol,
1: 1)) entering at the top of the middle section. The interaction between the solvents and
exhaust is encouraged on the glass tubing, which are packed into the middle tower section. The
extracted organics are removed, in the solvents, from the base of the tower.
21
,....,..----'1 L External cooling coils-------+..:.! . . ~
I
I
:1----r\- .-z-: I
I I
Baffies ------------t-';-t=;-o;;·\;-, -'"--if-"--
;: . -~,~
Insulation ----------t. ;:
I
: \ I _,_----'.._..._ T I
'-~~ Internal cooling coils-------+-·~ re; ::.-::,
c:e::==~
Solvent input ----------t---·lr SOLVENTS
'i! EXHAUST
Graded glass tubing -------~
rc::2 c:: c::::l c_] ----------~--·r o \ Heat shield
1 Exhaust input o
0 -
. I . I
~
T -s '0 0 -z 0
6 UJ Cl)
Q.. 0 E-
~ 6 UJ Cl)
UJ ...J 0 0 ~
+ ;:EZ oOe ~~='<t OUN Cll~o
Coolant -----------€=~1--__..._,,___t~--Solvent output 11
Figure 2.2 The total exhaust solvent-stripping apparatus (TESSA) . (Trier 1988)
22
The tower is cooled at the top and bottom to reduce post-combustion reactions and enhance
the condensation of the solvents. The system has a higher collection efficiency of gaseous
organics than is possible with filters alone, and combines rapid sampling with low artefacts
(Trier et al. 1988 and Rhead & Trier 1992).
The system devised by Kruzel et al. (1991) was similar in concept to TESSA, in that it
extracted raw exhaust by passing the exhaust through three progressively cooled flasks
containing DCM. The system designed for biological testing has the advantage of extracting
a large organic mass rapidly combined with low artefact potentials, due to the very low
collection temperatures (down to as low as -70°C in the last flask). The disadvantages of the
system for rapid profiling work is the complexity of the work-up for the large volumes of
DCM involved.
2.1.2 Laboratory Work-up & Analytical Techniques for Or.gauic Compounds
Following collection, samples require processing before the required chemical information of
interest can be obtained. The degree of laboratory work-up and sophistication of the detection
systems depend on the level of the analytes in the extracts; trace components require
complicated work-up and sensitive/selective detectors. The majority of chemical
characterization procedures follow an extraction (Section 2.1.2.1), fractionation (Section
2.1.2.2), and detection (Section 2.1.2.3) pathway.
2 1 2 1 Extraction and Concentration of Organics
Following sampling, the organics collected on filters and adsorbent traps require extraction.
The most common extraction techniques are those of soxhlet and ultrasonic extraction, with
the former being more effective and the latter considerably quicker (6-8 hours compared with
30 minutes)(Williams 1990). Dichloromethane is an ideal candidate for such extractions, since
it extracts 97% of the mutagenicity of diesel particles (Montreuil et al1992). The extracted
samples are termed the solvent-soluble organic fractions (SOF).
Engine sampling utilizing solvents as the collection medium dispense with soxhlet and
ultrasonication extractions, and in the case of TESSA the samples are termed the TESSA
extracted samples (TES).
23
2 1 2 2 Fractionation of Organics
Following extraction and concentration, the more prominent compounds, such as the n-alkanes
and two to four ring PAH, can be analyzed directly after extraction and concentration (Farrar
Khan et al. 1992). For detailed characterisation, the extract is fractionated into different
chemical classes. There are many ways to achieve fractionations, with the degree of isolation
required determining the complexity of the method. For compounds prominent in the extracts,
fractionation commonly proceeds by employing silica gel packed into open glass columns (Yu
& Rites 1981 and Williams et al. 1986) or solid-phase extraction cartridges (May & Wise
1984, Obuchi et al. 1984, Robbat et al. 1986, Theobald 1988, and Bundt et al. 1991). Other
fractionation methods include liquid-liquid partitioning, for example with dimethylsulphoxide
(Lee et al. 1981 and Henderson et al. 1984).
The analysis of minor exhaust constituents, such as nitro-PAC, commonly uses multi-step
fractionation, involving a simple preliminary clean-up step, followed by semi-preparative
normal-phase (NP) high-performance liquid chromatography (HPLC) (Levine & Skewes 1982,
Liberti et al. 1984, Tong et al. 1984, Ciccioli et al. 1986, and Bayona et al. 1988). The
complexity of HPLC fractionations are offset by isolating the specific elution window
associated with the analytes, allowing a much greater concentration step.
2 1.2.3 Detection of Organics
End analysis for organics is generally based on chromatographical techniques, and can be
divided between HPLC and gas chromatographic (GC) systems. Chromatography is used to
separate components of a sample, in which the components are distributed between two phases,
one phase is stationary, whilst the other moves (Braithwaite & Smith 1985). As the compounds
are carried over the mobile phase, there will be competition between the affinity to remain
with the stationary phase and carry on with the mobile phase (Figure 2.3). Since, different
compounds have varying distribution coefficients• between the two phases, separation of the
mixture occurs.
1 Distribution coefficient (D) = [stationary phase]/[mobile phase]
24
InitaJ stage
(a)
Flow of mobile phase
/--~c~fflr~---------Intermedaate )\
stages "\----------.AAAA<"77"i""""'----ccc~
Final stage
nn CCCC
8888 ~ ~ l)etector
1 (b)
Concentration signal
A
B
c
~-~-------------
to RA tRB t RC
Retention Time (minutes)
Key A,B, & C = model compounds
t 0 = Time for mobile phase to pass through system t RA = Retention time of A
tRB = Retention time of B
tRC = Retention time of C
hA = Peak height for A h8 = Peak height for B h c = Peak height for C t w = Width at base of peak ~ = Peak width at Y1. height
Figure 2.3 An idealised example of the chromatography for three model compounds (a) and the resulting chromatogram (b) (Braithwaite & Smith 1985)
25
The larger the distribution coefficient, the greater period of time will elapse before the
compound is eluted, termed retention time, compared with a compound which has a smaller
distribution coefficient. Once, the separation has occurred the individual components can be
detected. The detection of the major constituents of diesel organic extracts, such as PAR and
n-alkanes, are well established (fable 2.1).
Table 2.1 Detection systems for n-alkanes and major PAR
Detection system Advantages Disadvantages References
Reverse phase (RP)-HPLC Fast analysis Resolution and Maher et al. (1989) with UV detection times, simple set- alkanes
up undetected
RP-HPLC with fluorescence Greater selectivity Same as UV Hansen et al. (1991), than with UV Ganchanja (1993), and
Williams et al. (1994)
GC with flame ionisation High resolution, Low sensitivity Yu & Rites (1981), Niles detection (FID) stable, large linear and non-selective & Tan (1989), Bundt et al.
range (1991) , and Castello & Gerbino (1993)
Gas chromatography/mass Improved Greater Rites (1989) and Farrar-spectrometry (GC/MS) in identification complexity Khan et al. (1992) electron impact (El) mode
Minor constituents, such as nitro-PAC, require more selective and sensitive chromatographic
systems (fable 2.2). Higher sensitivities can be achieved for the fluorescence and
chemiluminescence detection if the nitro- group is reduced to the amino moiety. This can be
achieved on-line with the HPLC system (fejada et al. 1982, MacCrehan et al. 1988, and Liu
& Robbat 1991), or off-line (Campbell & Lee 1984, Sellstrom et al. 1987, and Hayawaka et
al. 1992). Derivatization techniques can be used to increase the selectivity and sensitivity of
ECD, and GC/MS in both El and NICI modes (Campbell & Lee 1984, Sellstrom 1987, and
Scheepers et al. 1994). Selective ion monitoring also greatly increases sensitivity for GC/MS
in El and NICI modes (Ramdahl & Urdal 1982, D'Agostino et al. 1983, and Schneider et al.
1990). Gas chromatographic analysis, offers the greatest resolution.
26
Table 2.2 Detection systems for nitro-PAC
Dete<:tion system Advantages Disadvantages References
RP-HPLC with fluorescence Speed and Resolution MacCrehan et aL (1988),
detection sele<:tive Tejada et al. (1986), Liu & Robbat 1991, and Veigl er al. (1994)
RP-HPLC with Sensitive and Resolution Sigvardson & Birks (1984),
chemiluminescence detection very selective Liu & Robbat 1991 , and Li & Westerholm (1994)
RP-HPLC with Sensitive and Interferences Rappaport et al. (1982), Jin e1ectrochemical detector large linear range from carbonyl & Rappaport (1982), and
compounds Galceran & Moyano ( 1993)
GC with electron capture Sensitive & good Limited linear Oehme et al. (1982), detection (ECD) resolution range & prone Campbell & Lee (1984), and
to contamination Draper (1986)
GC with nitrogen/phosphorus Highly selective Baseline D 'Agostino et al. (1983), detection (NPD) unstable Paputa-Peck et al. (1983),
White et al. (1984) , and Fetzer (1989)
GC with thennal energy Highly selective, Prone to peak Tomkins et al. (1984) and Yu analyzer (TEA) similar responses tailing & et al. (1984)
thennal degradations
GC/MS in El mode Distinctive Poor sensitivity Schuetzle et al. (1982), Tong spectra et al. (1984), and Ramdahl et
al. (1985)
GC/MS in negative ion Sele<:tive and Limited spectnt Oehme et al. (1982), Newton chemical ionisation (NICI) highly sensitive. et al. (1982), Ramdahl & mode Provides Urdal (1982), and Bayona et
mole<:ular ion al. (1988) data
Mass spectrometry/mass Highly sele<:tive Complex Henderson et al. (1982) and spectrometry (MS/MS) instrumentation Schuetzle et al. (1982)
However, the GC injection systems must be kept well maintained to prevent thermal
decomposition, and hydrogen gas is not suitable as a carrier due to possible reductions of nitro
PAC to amino-PAC in its presence (Tong et al. 1983 and White 1985). Liquid
chromatographic analysis have a faster sample throughput and do not suffer from the thermal
27
problems associated with some GC detection systems. Increasingly HPLC systems are the
being used to detect nitro-PAC, especially those employing fluorescence/chemiluminescence
detection with on-line reduction (Li & Westerholm 1994 and Veigl et al. 1994). The technique
of LC/MS is increasingly being used to identify the polar PAC present in diesel extracts
(Galceran & Moyano 1994).
2.2 Review of Previous Profiling Studies
The review is primarily focused on naturally aspirated light-duty DI engines (capacity no
greater than 2.5 I in total), however important findings from turbo, IDI and medium- and
heavy-duty engine research are in some cases reported. The section is sub-divided into
reviewing specific classes of organics, namely PAH (Section 2.2.1), alkanes (Section 2.2.2),
nitro-PAC (Section 2.2.3), and oxy-PAC (Section 2.2.4).
The majority of the reported studies utilized dilution tunnel and filter collection engine
sampling. As a consequence the chapter removes the engine sampling apparatus details, unless
the dilution engine sampling technique affected the results or a different engine sampling
system was employed. Details of the work-up and analytical techniques are also removed
unless relevant to the extent to which the profiling progressed.
2.2.1 PAH Emission Profiling
The early investigations into the effect of engine conditions on organic emissions focused on
PAH emissions, since compounds such as benzo(a)pyrene, were known carcinogens. The early
investigations tended to compare SI engines with diesels, and IDI with Dl fuel systems. The
affect of fuel and oil compositions on the exhaust make-up was largely unconsidered.
Williams & Swarin (1979) compared the benzo(a)pyrene emissions from diesel engines with
non-catalyst petrol powered engines. The averaged value of 2. 7 ILg/mile (ranging from 0.29
ILg/mile to 5.3 ILg/mile for different vehicles) from the SI engines was similar to the 2.7
ILg/mile emission rate from the diesel.
28
Kraft & Lies (1981) sampled four light-duty diesels (1978 Models: Rabbit & Golf, 1979
Models: Passat & Audi) over the FTP, highway fuel economy test (HWFET) and the
European regulation (ECE) driving cycles. The authors found the PAH emissions were lower
for the hot start HWFET cycle compared with the cold start FTP and ECE cycles.
Zierock et al. (1983) compared PAH emissions from a 3.8 I DI engine to those from a 1.5 I
IDI engine. The authors found the two injection configurations resulted in different emission
patterns over a wide range of speeds and loads (Figure 2.4). The DI engine produced peak
PAH concentrations at high and low loads with low speeds, whereas emissions were elevated
at high speed and across the load range for the IDI.
The study by Andrews et al. (1983) of a light-duty 2 I DI Petter AVl diesel engine was one
of the first studies to correlate the PAH in the exhaust to the PAH fuel content. The authors
found that for a speed of 1500 rpm, the weight of emissions relative to the mass of fuel
burned, decreased with increased load. For example, the emission of phenanthrene at low,
mid, and high loads was 0.15 mg/kg, 0.06 mg/kg, and 0.04 mg/kg respectively. This suggests
that the greatest combustion efficiency occurs at high loads for DJ engines, and consequently
that Dls are susceptible to poor combustion under low loads.
Jensen & Hites (1983) examined the effect of load (zero, half, and full) and injection timing
on PAH and alkyl-PAH emissions for a Cummins single cylinder heavy-duty diesel. Injection
timing did not influence the emissions; whereas load was shown to be inversely proportional
to emissions. Hence, zero load produced the greatest concentration of PAH and alkyi-PAH
emissions relative to the mass of particulate produced. Since the exhaust temperatures were
found to be proportional to engine load, the results indicated that the emissions were greatest
under low temperatures.
Mills et al. (1984) studied the effect of speed and load on the PAH emissions from a two
cylinder light-duty Petter BA2 DI 1.2 1 engine and expressed the emissions in terms of the
mass of PAH emitted relative to the exhaust volume {ltg/m3). The authors found at 2350 rpm
the total PAH were high at low loads, reduced at mid load, and further increased once more
at higher loads. Re-expressing the results relative to the fuel consumed, shows the highest
PAH emissions associated with low loads.
29
(a) IDI (b) DI
- e 260 El 120 c. c. .e. .e. = 200 Cl 100 Q Q ·.c ·.c 800
~ ~ loo loo 1273 -- = = ~ ~ 600 g 100 273.1 Cl Q 400 u u = = 200 < < ~ Speed (rpm) ~ Speed (rpm)
Load (bar) Load (bar)
Figure 2.4 The effect of engine speed and load on the PAH emissions from IDI (a) and D I (b) diesels . (Zierock et al. 1983)
30
Shore (1984) reported that for light-duty diesels, the mass of emissions per unit time were a
function of speed and load. At high and intermediate speed pyrene emissions decreased with
raised loads, whereas at lower speeds, pyrene increased directly with load. Shore (1986)
quantified the PAH emissions from a single cylinder Ricardo Hydra light-duty engine using
DJ and IDI versions, and for full and light loads with a selection of different fuel blends. For
US standard diesel, the DJ version emitted higher levels of benw(a)pyrene at light load (5.2
~tg/hour) compared with full load (2.4 ~tg/hour). In contrast, the IDJ version produced the
highest emissions at full load (4.6~tg/hour) relative to light load (3.3 ~tg/hour). The ratio of
P AH recovered in the exhaust as a percentage of P AH in the fuel supplied to the chamber were
higher for the DJ at full load compared with the IDI, for the US standard fuel. For example,
the recovery of chrysene with the DJ was 0.36 % whilst only 0.05 % for the IDJ. However,
different fuels produced not only different recovery levels but also in some cases a reversal in
trends for the DJ and IDI comparison. The combustion of one fuel, very low in PAH,
produced a exhaust recovery of 2351 % for benz( a )anthracene for the DI version and 5558 %
for the IDI. This showed that sources other than fuel survival, most probably combustion
generation, were responsible for the PAH emissions from that particular fuel blend.
Williams et al. (1986) sampled a Petter A V 1 DJ 2 1 diesel at 1500 rpm for low, mid, and high
loads. The reported survival of major PAH and alkyl-PAH, expressed as the percentage weight
of PAC recovered in the exhaust relative to the weight of PAH supplied to the chamber, was
significantly greater at low load. In contrast to Andrews et al. (1983), the survival of some
PAH increased at high load relative to the mid-load position (Figure 2.5).
The two studies by Barbella et al. (1988 & 1989) investigated the effect of engine load on the
PAH emissions at a speed of 2000 rpm. The emissions were expressed relative to the mass of
fuel burned. The first study investigated the effect of four different fuels on the emissions from
a heavy-duty turbocharged Dl. For each fuel the engine was sampled at high and low air: fuel
ratios (a) (Table 2.3). Generally, the PAH emissions were highest at low load (high a), except
for fuels C & D, where the trend in some cases was reversed.
31
'Gl c.:
1 Key
c • Fluoranthene ·-= < =- • Methyl-fluorenes - D Phenanthrene 0
~ "CC 0 Methyl-phenanthrenes ~
.~ ... Pyrene e ~
= < =--0 C)l)
e
High Mid Low
Engine Load
Figure 2.5 The effect of engine load on the PAH and Alkyl-PAR emissions from a light-duty D [ diesel. (Will iams et al. 1986)
32
Table 2.3 PAH Emission Rates (mg/kg) in DI exhausts at two different air: fuel ratios and for four fuel blends
Fuels A B c D
Air: fuel o:=23 ratio (a:) 11=44 11=2 11=37 u=2 u=45 u=23 11 =35
Naphthalene 4.4 9.7 0.4 31.5 1.3 1.8 0.6 -
Acenaohthvlene 7.0 10.0 8.5 44.6 11.4 7.5 4.3 8.3
Fluorene 2.6 4.5 1.0 14.0 2.4 1.3 8.8 0.6
Phenanthrene 7.0 19.0 3.7 42.2 6.3 3.5 2.1 3.1
Fluoranthene 0.4 0.5 0.2 0.5 4.9 - 0.1 -
Pvrene 0.6 2.5 0.3 7.6 0.5 - 0.2 -
Key to fuels (cetane No., % aromatic, fuU boiling point•C):
A 35.0, 56.0, 358 C : 49.5, 23.7, 370 B 35.0, 54.5, >400 D : 51.0, 34.8, 388 (Barbella et al. 1988)
In the second study Barbella et al. (1989) found the highest mass of PAH emissions relative
to the total fuel consumed, at low loads from a light-duty 2 1 DJ diesel. The emission rates of
fluorene and phenanthrene then decreased at mid-load, before rising once again at high load.
In contrast, the emission rates for fluoranthene and benzo(e)pyrene progressively declined with
increased load.
Williams et al. (1989) compared the organic emissions from an old Petter DI diesel to a
modern medium-duty Perkins DJ 4-236 4 1 engine. The authors found that the superior
technology incorporated into the Perkins engine corresponded to higher combustion
efficiencies. In contrast to other DJ studies, the Survival of 2-3 ring PAH from the Perkins
engine increased with elevated load at 2600 rpm. For example, the reported survival of
phenanthrene at low, mid, and high load was 0.05 %, 0.03 %, and 0.58 % respectively. The
different trends may reflect the larger engine size compared with the light-duty studies. The
study also found survival of some PAH were greater at increased speed, whereas other PAH
tended to follow a moderate decline with increased speed (Figure 2.6).
33
2.0
CV > 1.5 > L. :::l
(/)
~ 1.0 CV ..., c: Cl)
0 L. Cl) 0.5
a..
1000
o Phenanthrene
* !). Fluoranthene
1:!>~• * Pyrene
--------.
2000
Speed (rpm)
3000
Figure 2.6 The effect of engine speed on the survival of PAH fro m a medium-duty 01 diesel. (Williams et al. l989)
34
As noted by the authors the higher molecular weight PAH, such as benzo(a)pyrene, appeared
to be the result of significant combustion generation reactions, as reflected by the 60 % and
12 % 'survival' corresponding to 1200 rpm and 2600 rpm respectively.
Trier et al. (1990) compared the emissions from an IDI Ricardo E.6/T light-duty diesel at
2000 rpm and 3.5 kW with the emissions at 2750 rpm and 6.6 kW. The engine was sampled
using the unique TESSA developed at Plymouth University (Section 2.1.1). The concentration
of PAC in TES were expressed relative to those in the fuel, such that the GC/MS areas of
naphthalene in TES was divided by the area of naphthalene in the fuel after correcting for
injection volumes (all samples diluted to the same concentrations). In this way the dominance
of fuel survival and combustion chamber contributions to TES could be assessed for the two
engine powers. The results showed that the alkyl-PAC in TES relative to the alkyl-PAC in the
fuel were lower compared with the corresponding unsubstituted PAC ratios (Figure 2.7). Trier
et al. argued that the severity of the difference between the PAC and methyl derivatives at the
high power (fES 2) was strong evidence for pyrosynthetic reactions by which the unsubstituted
PAC were produced relative to the methyl derivatives.
Abbass et al. (199la) found the mass of PAH emissions relative to the mass of fuel burned at
2200 rpm increased dramatically under high loads, for a 0.22 I IDI Petter AAl diesel. The
reported smvival trend for the IDI contrast those found for light-duty DI engines, and reflects
the differing engine conditions under which the two injection systems are prone to combustion
problems.
Zeijewski et al. (1991) studied the effect of speed and load on emissions at twelve sampling
points, for a light-duty 2 I DI engine. The authors determined that unbumt fuel, speed and load
were the variables controlling the emissions. The highest emissions, expressed relative to the
total mass of fuel consumed, corresponded to low loads. The highest recovery of exhaust PAH
occurred for 1100 rpm and under idling. At 2200 rpm, the PAH survival was highest at low
load, reduced at mid-load, and elevated once again at higher loads. At 3000 rpm, the survival
trend with load was replicated, only with a much smaller increase in the survival at high loads
relative to mid-load. The optimum speed for the greatest combustion was found to be that of
2200 rpm.
35
-Cl)
::s ~ c: ·-~ Cl) ;.... ~
~ ~ Cl) e::
C/)
gg c: ~ Cl) ;.... ~
~ ~ Cl)
~
·-
3
2
1
0
Key
NP =Naphthalene PR = Phenanthrene
C 1- = Monomethyl-
DUnsubstituted PAC
•Methyl-PAC - ... ........... -·
- .. -· - . ... '"'' . ....
I I
FR = Fluorene DBT = Dibenzothiophene
C2- = Dimethyl-
TES 1 .. . -·-
.. . .. ··-· .. ...... ..........
NP Cl-NPs C2-NPs FR Cl -FRs PR Cl-PRs DBT Cl-DBTs C2-DBTs
3 .-----------------------------------------------------,
2
1
DU nsubstituted PAC
• Methyl-PAC TES2
NP Cl -NPs C2-NPs FR C 1-FRs PR Cl -PRs DBT Cl-DBTs C2-DBTs
Figure 2.7 The ratio between PAC in TES relative to the PAC in the fuel for a) low engine power and b) high engine power. (Trier et al. 1990)
36
2 2 2 n-Aikane Emjssjon Profiling
The low mutagenic and carcinogenic nature of n-alkane emissions has resulted in less research
compared with the PAH emissions from diesel engines. The n-alkanes are also relatively
unreactive in the atmosphere (Boone & Macias 1987). Where n-alkanes have been investigated,
it was to assess the sources of the organic emissions over varying combustion conditions. The
simple profile of the homologous series of n-alkanes present in diesel fuel enables the survival
and combustion generation pathways to be investigated.
Williams et al. (1989) and Abbass et al. (1991a), reported in the PAH review (Section 2.1),
also investigated the effect of engine conditions on the combustion efficiency of selected n
alkanes. In the first study, the n-alkanes survived greatest at high loads for the 4 l medium
duty DI engine, reproducing the trend found for PAH. Survival for equivalent boiling point
P AH and n-alkanes, such as nonadecane relative to phenanthrene, were of similar magnitudes.
The higher relative molecular mass n-alkanes survived to a greater extent than lighter n
alkanes, taken by the authors to indicate a lower combustibility associated with larger chain
structures.
Abbass et al. (1991a) found that for the small 0.22 l IDI the highest survival of n-alkanes was,
as for PAH emissions, associated with high loads (Figure 2. 8). The survival comparisons of
equivalent n-alkane and PAH boiling point compounds, were not as close as those found by
Williams et al. (1989), with the PAH survival higher; especially for large PAH. The authors
suggest this may be evidence for combustion chamber generation of PAH. However, as the
authors note, the role of engine/exhaust deposits in producing the observed emissions could
not be ruled out.
37
0.00 l+r---,----T-----r--__,..--,--r---,----,---_,....---r----. 10 14 18 22 26 30
.n-Alkane carbon number
Figure 2.8 The effect of engine load on the survival of n-alkanes emissions from a lightduty IDI diesel. (Abbass et al. 199la)
38
2 2 3 Nitro=PAC Emission Profiling
The discovery in the late 1970s that nitro-PAC emissions were present in diesel extracts,
invoked great interest due to the high mutagenic potency of some nitro-PAC, such as 1-
nitropyrene. The majority of research to date has been directed at developing fractionation and
detection methods capable of detecting the trace levels of nitro-PAC in the exhaust samples
(Jiiger 1978, Xu et al. 1981 & 1982, Newton et al. 1982, Ramdahl & Urdal 1982, Rappaport
et al. 1982, Hanson et al. 1983, Jin & Rappaport 1983, Nakagawa et al. 1983, Tong et al.
1983, Brown & Poole 1984, Chou 1984, Sigvardson & Birks 1984, Tomkins et al. 1984,
White et al. 1984, Yu et al. 1984, Lindner et al. 1985, Ramdahl et al. 1985, Wise et al.
1985, Ciccioli et al. 1986, La Course & Jensen 1986, Robbat et al. 1986, Sellstrom et al.
1987, Bayona et al. 1988, MacCrehan et al. 1988, Niles & Tan 1989, Imaizumi et al. 1990,
Schneider et al. 1990, Liu & Robbat 1991, Hayakawa et al. 1992, Fu et al. 1993, and Li &
Westerholm 1994).
The profiling of oitro-PAC in diesel exhaust over a range of speeds and loads, resulting in the
identification of engine conditions which favour nitration reactions is largely uncharted.
Research so far has tended to establish the levels of selected nitro-PAC at a specific engine
condition (Oehme et al. 1982, Schuetzle et al. 1982, Paputa-Peck et al. 1983, Campbell &
Lee 1984, Hartung et al. 1984, and Liberti et al. 1984); and there has been little research
aimed at sampling a range of engine conditions for a particular engine. One of the reasons for
this lack of data is attributable to the technical instrumentation needed to detect the low levels
of nitro-PAC in diesel exhaust. The low levels require lengthy sampling periods before the
dilution tunnel and filter collection method can acquire sufficient organic material for analysis
to proceed (Bechtold et al. 1984). The long sampling periods have in some cases resulted in
artefact formation of nitro-PAC (Pitts et al. 1978, Gibson et al. 1981, Bradow et al. 1982,
Herr et al. 1982, Risby & Letz 1983, Lindskog 1983, Chan & Gibson 1985, Hartung et al.
1984, and Lach & Winckler 1988).
Gibson et al. (1981) examined the emissions of 1-nitropyrene and 6-nitrobenzo(a)pyrene from
General Motors diesel engines using the hot-start FTP. The older 1978 5. 7 l engines produced
significantly higher nitro-PAC emissions than did the 1980 5.71 engines.
39
Tejada et al. (1982) determined the concentration of 1-nitropyrene in the SOF from a 1978
Volkswagen Oldsmobile diesel and a 1980 Rabbit diesel on a selection of driving cycles, such
as the FTP, HWFET, and New York city cycle (NYCC). The Rabbit diesel produced less
emissions than did the Oldsmobile for all driving cycles. The highest 1-nitropyrene emissions
were for the NYCC, where the Rabbit produced 63.8 ppm and the Oldsmobile 111 ppm. The
levels of 1-nitropyrene emitted from the diesel engines were far greater than those from SI
engines tested on the same cycle. For example a Ford Mustang SI engine produced 1-
nitropyrene emissions of only 12.7 ppm over the NYCC. The authors also investigated the
nitration of pyrene by adding N02 to the exhaust, whereupon 1-nitropyrene increased at the
expense of pyrene, as the N02 level in the exhaust was progressively increased.
Schuetzle & Perez (1983) found the concentration of 1-nitropyrene in the SOF from a heavy
duty diesel operating at 2100 rpm, initially increased with load, before decreasing with further
load (Figure 2.9). For partially oxidized nitro-PAC, such as 3-nitrofluoren-9-one, the
concentrations increased with load at the same speed, as did the dinitropyrenes. The highest
emissions of 1-nitropyrene were found at low load and low speed, whilst the least emission
variability for 1-nitropyrene was associated with 1260 rpm. The authors found the emission
rate of 1-nitropyrene for light-duty diesels ( 4. 7 ILg/krn) was higher compared with heavy-duty
diesels (1.5 ILg/km) operated over the transient FTP cycle.
Bechtold et al. (1984) collected the exhaust from a 1980 5.71 eight cylinder Oldsmobile diesel
over segments of the FTP cycle (total cycle time of ea. 23 minutes). The sampled segments
included idling, accelerating periods, and steady state cruising sections. Repeated sampling was
needed to gain sufficient SOF (ea. 200 mg) for biological and chemical analysis, especially
in the case of the cold acceleration segment, where 33 consecutive cycles were combined. The
authors found a strong correlation between the mutagenicity and both the exhaust temperature
and gaseous exhaust NO, level. Chemical analysis indicated that 1-nitropyrene was also related
to NO, and possibly temperature (Figure 2.10).
40
- 30 E A 700 rpm a. a. 25 - * 1260 rpm ...., 0 «< .... 20 0 2100 ...., rpm >< Q)
c: 15 c: 0 ....,
10 «< .... -c * Q)
0 5 c: 0
(.) 0
0 20 40 60 80 100
Percentage of Full Load
Figure 2.9 The effect of engine conditions on 1-nitropyrene emissions from a heavy-duty diesel. (Schuetzle & Perez 1983)
41
E a. a.
c 0 -.. ...
80
60
c 40 Cl 0 c 0 0
- 20 0 .. .... -)( w 0
20
E a. 18
_g. c 16 .g :! 14 -c: Cl 0 12 c: 0 0
10 .. ::J 0 Cl ... .. "
8
6
350
~ 325
Cl ... z 300 fll .... ~ 275 E Cl ....
250 Cl a.
~ 225 .. 1-
200
a) 1-Nitropyrene
Idle 30 mph cruise 55 mph cruise
b) NOx
Idle 30 mph cruise 55 mph cruise
c) Temperature
Idle 30 mph cruin 55 mph cruise
Figure 2.10 The effect of different segments of a FTP cycle on the emiSSions of 1-nitropyrene (a), emissions of NOx (b) , and temperature (c). (Bechtold et al. 1984)
42
Handa et al. (1984) examined wall deposits from two tunnels for vehicle emissions, one on
an ascending gradient (causing the vehicles to be under high loads) and the other on a
descending gradient (vehicles subjected to a lower load burden). The vehicle flow rate was
automatically recorded and the vehicle type divided into two classes according to the vehicle
length: less than 6 m corresponded to mostly light-duty SI engines and greater than 6 m to
heavy-duty diesel vehicles. The heavy-duty diesel contribution to 1-nitropyrene emissions at
high load was 24 1-'g/hr compared with 6.6 1-'g/hr at low loads. The levels of 1-nitropyrene
emitted by diesels were ea. ten times greater than those from SI engines. The authors found
that the nitro-PAC/PAH ratio was much higher for diesels compared to SI engines, and
concluded that the nitro-PAC emissions were favoured at the high NO, and temperatures
associated with diesel combustion at high load.
Draper (1986) quantified the nitro-PAC emissions from a heavy-duty diesel engine at two
different engine conditions. The higher relative molecular mass nitro-PAC emissions were
greater at moderate load and high speed compared with high load and moderate speed (Table
2.4). The study is restricted by the limited sampling positions, and the elucidation of whether
speed or load is responsible for the nitro-PAC emissions is further complicated by both
variables changing for each sampling position.
Table 2.4 Effect of two different engine conditions on the Nitro-PAC Emissions from a heavyduty diesel
Compound High Load, Moderate Speed Moderate load, High Speed (pg/g) (j.tg/g)
1-nitronaphthalene 0.77 0.47
2-nitronaphthalene 0.94 0.87
2-nitrotluorene 0.63 8.8
9-nitroanthracene 0.34 1.4
1-nitropyrene not detected 5.0
1, 3-dinitropyrene 0.52 1.6
(Draper 1986)
43
Shore (1986) investigated the nitro-PAC emissions from Dl and IDI versions of a Ricardo
Hydra diesel, operated at 2400 rpm and 7 bar. For a standard US diesel fuel the Dl produced
substantially higher emissions of 1-nitropyrene (11.1 JLg/hour) compared with the IDI (0.78
JLg/hour). The NOx emission rate of the Dl was typically three to four types times that of the
IDI engine. Shore could find no correlation between the 1-nitropyrene emissions and different
fuel blends.
Hirakouchi, N. et al. (1990) analyzed the exhaust from a 6.5 I turbocharged Dl diesel and a
11.21 normally aspirated Dl diesel for nitro-PAC. The turbocharged diesel, sampled over the
FTP cycle, produced nitro-PAC emissions of 1.07 x let g/BHP-hour of 2-nitrofluorene, 3.29
X 10"6 g/BHP-hour of 9-nitroanthracene and 5.90 X 10-6 g/BHP-hour of 1-nitropyrene. The
11.2 I diesel was sampled at mid-speed for 25%, 50%, 75% and 100% of full load. The nitro
PAC emissions were expressed in terms of mass emitted per volume of gases produced (Figure
2.11). Different nitro-PAC produced different trends with respect to load. For instance
9-nitroanthracene and 2-nitrofluorene were highest at low load, reduced at mid-loads and
increased at high load. The emissions of 1-nitropyrene and 3-nitrofluoranthene progressively
decreased with increasing load, whereas the nitronaphthalenes were more stable across the load
range. The mass of 1-nitropyrene emitted relative to the total fuel consumed was highest at low
loads.
Scheepers et al. (1994) evaluated the 1-nitropyrene emissions from heavy-duty and light-duty
diesel engines over different driving cycles (Table 2.5). There were only small differences
between varying driving conditions for both engines. The oxidation catalyst fitted to the light
duty engine resulted in very low levels of 1-nitropyrene and particulates. In fact, after
cOnducting a series of tests, 1-nitropyrene was only detectable in three out of sixteen samples.
This highlights the inherent problem with dilution tunnel sampling for trace nitro-PAC
analysis, namely that of lengthy sampling periods for sufficient SOF collection, which can in
some cases lead to significant artefact contributions to the nitro-PAC.
44
0.010
0 .008
0 0.006 ('I")
E ......... C)
::L 0.004
o 1- Nitronaphthalene A 2-Nitronaphthalene v 2- Nitrofluorene 0 9- Nitroanthracene * 3-Nitrofluoranthene o 1- Nitropyrene
0.002 0-----1
A.------
20 40 60 80
Percentage of Full Load
100
Figure 2.11 Effect of engine load on the nitro-PAC emissions from a heavy-duty turbocharged diesel. (Hirakouchi et al. 1990)
45
Table 2.5 Effect of Driving Conditions on 1-Nitropyrene emissions from Light- and HeavyDuty diesels
Vehicle Driving Pattern 1-nitropyrene
ul!ll! narticulate Ul!lkm travelled
Heavv dutv Sub urban 1.94 2.13
Sub urban 1.94 3.28
Urban 3.39 6.48
Motorway 1.94 0.87
Motorwav 2.73 1.37
Lil!ht dutv1 European drivin11. cvcle (cold start) _2 0.32
European driving cycle (hot start) _2 0.22
US '75 Drivinl! cvcle _2 0.36
1 With oxidation catalyst 2 Amount of particulate matter not determined
(Scheepers et al. 1994)
Veigl et al. (1994) determined the emissions of 1-nitropyrene from a 1.8 I four cylinder
normal aspirated DI diesel and a 4 l four cylinder turbocharged DI diesel. The light-duty
engine was tested over the US FTP and HWFET cycles. The predominately high speed
HWFET cycle produced considerably higher levels of 1-nitropyrene compared with the more
varied FTP cycle. The authors also found that the addition of EGR and an oxidation catalyst
reduced the 1-nitropyrene emissions. The larger engine was sampled using steady state testing
at 1560 rpm and across the load range. The engine was also tested at very low speed and load
idling conditions (Figure 2.12). The highest 1-nitropyrene emissions for 1560 rpm were at
mid-load, whilst the lowest emissions were at 600 rpm and idling. The authors also found that
the addition of a fuel additive decreased the 1-nitropyrene emissions. The authors claim that
since the emissions are reduced by both engineering and fuel changes, artefacts are not
responsible for the results.
46
120 o 1560 rpm
- 11 600 rpm ~
::l 100 0 .c
' C)
I __ J ::t. - 80
""C Q) ..... ..... E 60
D 1 Q)
1 Q) c: D Q) ~ 40 > 0. 0 ~ ..... c:
20 I ......
0 20 40 60 80 100
Percentage of Full Load
Figure 2.12 The effect of engine load on the 1-nitropyrene emissions from a medium-duty Dl diesel. 0/eigl et al. 1994)
47
2.2.4 Oxy-PAC Emission Profiling
The realisation that nitro-PAC may not account for all of the direct acting mutagenicity of
exhaust extracts led to investigation of other possible candidates. Some oxy- and polar PAC
present in diesel exhaust have high mutagenic potencies (Scheepers & Bos 1992a).
The composition of the oxy- and polar PAC fractions of diesel exhaust consists of aldehyde,
ketone, and quinone derivatives of PAH (Yu & Hites 1981, Choudbury 1982, Ramdahl 1983,
Kooig et al. 1983, and Schulze et al. 1984). The analysis of the polar PAC in exhaust samples
by gas chromatography is hindered by the volatility problems, such as PAC carboxylic acids
(Oehme 1985). The complexity of the polar derivatives of PAH necessitates techniques, such
as LC/MS, to overcome the problems associated with GC analysis. As a consequence of the
technical problems, the determination of the effect of engine conditions on the oxidizing
potential of diesel combustion remains largely unanswered.
Jensen & Hites (1983) found the oxy-PAC emissions from a heavy-duty Cummins diesel at
zero, mid, and high loads exhibited varying trends depending on the structure (Figure 2.13).
For one set of oxy-PAC (naphthalenecarboxaldehydes, fluoren-9-one, and
bipheylcarboxaldehydes) the emissions initially increased with engine load and then decreased
thereafter. The second set (phenanthrenecarboxaldehydes and thioxanthen-9-one) decreased
progressively with increased engine load and the corresponding higher exhaust temperatures.
Schuetzle & Perez (1983) showed that partial oxidation of nitro-PAC may become more active
at high speeds and loads for a heavy-duty diesel. For example, the 17 ppm of 3-nitrofluoren-9-
one in the extract at low load increased to 63 ppm at high load. Similarly, 3-nitro-l ,8-
naphthalic acid anhydride was 22 ppm and 174 ppm at low and high loads respectively.
48
-0) 75 ........ 0)
:::1. -CD 60 .... tV :::::J 0
t: 45 tV a. ~
~
0 .... tV
30
!; 15 ~ CD 0 ~
0 0
0 20 40
o Naphthalenecarboxaldehydes A Fluoren-9-one v Biphenylcarboxaldehydes <> Phenanthrenecarboxaldehydes * Thioxanthen-9- one
60 80 100
Percentage of Full Load
Figure 2. 13 Effect of engine load on the oxy-PAC emissions from a heavy-duty diesel. (Jensen & Hites 1983)
49
The study of wall deposits on two tunnels (one inclined and the other declined with respect to
the traffic flow) by Handa et al. (1984) showed the hourly emissions of benz(a:)anthracene
quinones, benzo(a)pyrene quinones, perylene quinones, and benzanthrone increased at high
load. The oxy-PAC were also produced in greater quantities from the heavy-duty diesels
compared with the SI light-duty vehicles.
Hirakouchi et al. (1990) found that the aldehyde and ketone emissions, such as formaldehyde,
benzaldehyde, and acetone, relative to the total fuel consumed, were greatest under low loads
for a heavy-duty diesel.
2.3 Summary of Emission Profiling & Aims of Investigation
In terms of the technical aspects of emission profiling, the choice of the engine sampling
system is crucial to the type of profiling produced. Dilution tunnel/filter systems are not suited
to sampling secondary nitro- and oxy-PAC emissions, for which extended sampling leads to
high artefact contributions. The simulation of the initial atmospheric dilution of the diesel
exhaust by dilution tunnels, hinders the differentiation between the role of the exhaust and
combustion chamber on emissions. Sampling with TESSA allows the role of diesel combustion
in the formation of primary and secondary organic emissions to be investigated; information
vital to engineering and fuel companies alike. The TESSA enables large organic extracts to
be rapidly collected (sampling times of the order of seconds) with minimum artefact
contributions, thus allowing a wide range of speeds and loads to be profiled.
The profiling of the primary n-alkane and PAC emissions can be easily achieved with
relatively simple work-up and detection systems. The most informative profiling units are
those which relate the mass of a specific compound emitted relative to the same compound
within the fuel consumed during sampling. Such units enable the combustion efficiency of
different organics under varying engine conditions to be followed. The total mass of emissions
collected during driving cycle simulations are not as informative as steady state tests in
enabling the identification of engine conditions responsible for combustion problems. Light
duty DI diesels are prone to combustion problems at low loads. By contrast, the IDI diesel is
prone to becoming inefficient at high load, as are larger Dls, emphasizing the effect of engine
50
design on the combustion process. There is limited evidence for the combustion efficiency of
light-duty DI engines being greatest at mid-speeds. The n-alkane emission proftles closely
resemble those found for PAH, and consequently the highest survival of n-alkanes occurs at
engine conditions under which the greatest PAH survival occurs. To conclude, the role of the
combustion chamber on the primary emissions from light-duty DI diesels over a wide range
of conditions, especially speed, requires further research.
There are great sampling and analytical problems associated with profiling trace nitro- and
oxy-PAC emissions. With regard to the highly mutagenic nitro-PAC, there have been few
comprehensive studies on the effect of engine conditions on the nitration potential of diesel
combustion. Those studies which have been conducted relied on dilution tunnel sampling
systems, not suited to rapid mass sampling and prone to artefact formation of nitro-PAC. The
majority of the steady state sampling studies have involved heavy-duty engines, and the
similarity of heavy-duty and light-duty combustion in terms of nitro-PAC is largely unknown.
Two formation theories for nitro-PAC emissions from diesel engines have been proposed:
1) Combustion generation involving PAH and NO •.
2) Post-combustion electrophilic nitrations between PAH and N02 (in the
presence of nitric acid). (Scheepers & Bos 1992b)
The first pathway may be linked to the radical processes involved with high temperature
combustion, such that PAH combine with N02 and NO radicals, followed in the case of NO
by further oxidation (Nielsen et al. 1983). The second route would proceed under acidic
conditions, by the action of the nitronium ion (N02 +), as is commonly employed in laboratory
synthesis of nitro-PAC (Nielsen 1984 and Ross et al. 1988). It is important to stress that
neither theory has been proved under specific engine conditions. This is due to dilution
tunnel/filter sampling systems being unable to differentiate between the combustion and post
combustion contributions to the sample matrix. To conclude, there is insufficient information
to enable elucidation of the combustion factors controlling the nitration of primary PAH
emissions from light-duty DI diesel engines.
51
In view of the limited information on the highly hazardous secondary nitro-PAC emissions,
the aim of this study was to achieve comprehensive profiling for such emissions using TESSA
sampling close to the combustion chamber for a light-duty DI diesel. The objectives of the
research can be summarised as follows:
1) To establish the relationship between the primary emissions present in the fuel, such as
PAH, with engine speed and load over a range of conditions. To use the engine map to
understand under which conditions the greatest fuel survival and combustion generation
contributions to the emissions are favoured.
2) To develop and validate laboratory work-up and analytical techniques capable of detecting
the secondary nitro-PAC emissions.
3) To employ the developed analytical routines to produce a comprehensive engine map of
nitro-PAC emissions from the engine. To use the nitro-PAC map to examine the effect of
speed and load on the nitration of the primary PAH emissions, and from this postulate
combustion/exhaust areas to which the two proposed nitration mechanisms may be favoured.
4) To draw conclusions from the primary and secondary emission maps, which may be of use
to both the engineer and fuel chemist, concerning the possible control strategies which could
be employed, if environmental pressure resulted in PAC legislation. Finally, to indicate where
further research projects could be directed.
52
Chapter 3 - Experimental Procedures & Facilities
This chapter details the sampling and analysis procedures involved in this research programme.
There are four main experimental sections, the first deals with the test facilities and the fuel/oil
used throughout the study (Section 3.1). The profiling of the primary emissions (PAH and n
alkanes) over a range of speeds and loads using the initial TESSA configuration is described
in Section 3.2. Following the first profiling research, TESSA was relocated, upgraded and
tested (Section 3.3). The second profiling research package investigated the secondary
emissions (nitro- and oxy-PAC) over a range of engine conditions using the upgraded TESSA
(Section 3.4).
3.1 Facilities and Fuel/Oil Description
This section covers the hardware required for engine sampling; namely the engine itself
(Section 3.1.1), the engine management system (Section 3.1.2), and the engine sampling
system (Section 3.1.3). The type of fuel and oil used throughout the study are detailed in
Section 3.1.4.
3 1 1 Engine Details
The test engine throughout the research was that of a Perkins Prima direct injection light-duty
engine. The engine is fitted with a Bosch EPVE rotary fuel injector pump along with CA V 4-
hole fuel injectors. Full details are given in Table 3 .1. The engine exhaust manifold had
previously been modified to allow sampling from one cylinder (Rhead et al. 1991).
3.1.2 Engine Management
The test engine was coupled to a swinging field electric Borgi & Saveri FAlOOO eddy current
dynamometer managed by a Test Automation 2000 C-E compact controller, operated in the
constant torque mode.
3.1. 3 Total Exhaust Stripping Solvent Apparatus
This section provides technical details of the engine sampling system used for this research,
termed the total exhaust solvent-stripping apparatus (TESSA). The advantages of directly
sampling the exhaust close to the combustion chamber with TESSA compared with
53
conventional dilution and filter engine sampling are outlined in Section 2.1. This section
describes the configuration of TESSA at the beginning of the research.
Table 3.1 Perkins Prima Engine Details
Specification
Powerat4500rvm~~ 46
Toraue at 2500 mm (Nm) 122
Bore (mm) 84.4
Stroke (mm) 88.9
Number of cvlinders Four in-line vertical
Caoacitv (]) 1.993
Cvcle 4 stroke
Asoiration Natural
Combustion svstem Prima one-stage direct injection
Comoression Ratio 18.1: I
The system consists of a stainless steel tower into which the exhaust is fed directly from one
combustion chamber of the Prima engine. The organics are removed from the tower by a
downward flow of solvent mixture (DCM:methanol, 1:1). The tower is divided into three
sections (Figure 3.la). The cooled bottom section incorporates the exhaust entry point. The
exhaust entry point is connected to the engine exhaust manifold by a short transfer pipe. The
solvent flow is set by tap 1 and sampling of the exhaust proceeds by releasing tap 2, allowing
the solvents to enter the middle section. The exhaust is directed from the main exhaust into
TESSA by a slidable plate mechanism (Figure 3.lb). The middle section is filled with short
sections of glass tubes. The surface of the tubing provides an interaction point for the solvent
and the organics in the exhaust. The top section contains the exhaust baffles, and also provides
further cooling. The entire tower length can be washed down with the solvent mixture from
the entry point in the top section using tap 3.
54
Solvent reservoir
Tap 3
Top section
Bottom section
Tap 2
+- Exhaust from engine
1 Solvents and organics
(a)
Exhaust ducts
/ " To TESSA +- -+ To main exhaust
Exhaust duct from cylinder
(b)
Figure 3. 1 The operation of the solvent delivery system (a) and exhaust transfer mechanism (b) used for engine sampling with TESSA.
55
3.1.4 Fuels and OUs
Throughout the research a reliable fuel stock (1600 1) of standard No. 2 DERV (Table 3.2)
was purchased and stored in eight 50 gallon (ea. 225 1) drums. The drums, which previously
contained oil or industrial solvents, were thoroughly washed out with a portion of the standard
fuel many times before the drums were filled direct from the tanker. The engine was supplied
with fuel from a 10 1 head tank.
The head tank output entered a fuel/water separator before entering a 400 ml fuel plinth (with
measurement increments of 100 m!, 200 ml, and 400 ml). The standard return route for the
fuel system was to bypass the fuel plinth. A switch mechanism was installed to allow the fuel
supply to be isolated to the fuel plinth for fuel consumption measurements. The lubricating oil
used throughout the study was that of Shell Heavy Duty Rimula X, grade 15/30 W.
T bl 3 2 a e tan ar 0. tese ue roJerties S d d N 2 D' IF 1 P
Specification
Aromatic content (lP 156) 29%
Cetane index {lP 218) 1 49
Specific gravity (@ 60°F) 0.863
Viscosity (@ 40°C) 3.58 cS
Initial boiling point ea. lSS•c
5% 211•c
10% 231"C
20 % 2so•c
30 % 264•c
40% 21s•c
50 % 2s6•c
60 % 29s•c
70 % JJO•c
80 % 32J•c
90% 343•c
Final boiling point J64•c
1 Based on mid-boiling point and an API gravity of 32.5 Testing canied out by Perkins Technology
56
3.2 Profiling of the Primacy Emissions using the initial TESSA configuration
The profiling of the primary emissions (PAH, PASH and n-alkanes) from the Prima engine
was attempted with the initial TESSA set-up. The overall procedure for the sampling, work-up
ofTES, and analysis of specific organic classes is shown in Figure 3.2. The sampling and the
work-up details forTES are given in Sections 3.2.1 and 3.2.2 respectively. The analysis of
the n-alkanes GC-FID is detailed in Section 3.2.3. The analysis of the PAH and PASH by
GC/MS operated in the El mode is described in Section 3.2.4.
3 2.1 Engine Sampling for Primary Emissions
Engine sampling was performed in series, referring to sampling at a constant speed and at
different percentages of full load positions. The full load for a particular speed was arrived at
by increasing the load until the highest stable load was obtained. Following the acquirement
of the full load position, 1%, 25%, 50%, 75% of full load were calculated.
Prior to engine sampling the chiller was filled with water/ice and circulated around the cooling
coils of the tower. The engine was conditioned for one hour at 3000 rpm and 90 Nm, followed
by further conditioning for fifteen minutes at the desired engine conditions. The solvent
reservoir was filled with the solvent mixture (800 ml) and the reservoir opened (tap 1) to the
set position. The solvent flow into the top of the middle section of the tower was started (tap
2) ten seconds before switching the exhaust from the main exhaust into the tower by the
manually operated sliding plate. The exhaust was sampled for thirty seconds before switching
the exhaust back to the main outlet. The solvent flow was directed to the top of the tower (tap
3}, and the tower washed down with the remaining solvent.
The engine was then conditioned for the next load position, and the sampling procedure
repeated until all the desired load positions had been sampled on the same day. For each
engine series a fuel and sump oil sample were taken. Three engine series were completed at
1000 rpm, 2000 rpm, and 3000 rpm on different days (Tables 3.3- 3.5).
57
Figure 3.2
Prima sampled u.ilh lESS..-\
I Int .. nal Standards added
I
.·.·-:-·-:-:-:-·-:-·-·-:-:-:-:-:-:;:·-:-·-·-:-:-·-:-·-·-·-:-::-::-:-:-·-:-:-:-·-.
I
..-.·.:-·-:-·-·.·.·-·-:::-:.·· -·--··.· -·-:-·-·-::-;-·-·.·.-.·.
Ana lytical K ey:
GC-FID = Gas d tro m a to graph y with tl ruu e ionisatio n dete~tion
GC/MS El = Gas .;hromatgraph y I mass speclromelry operated in e lectron impact tnode
l I I
.-\liplllltia .-\ranatia Polan
--~---.·· ··.·· ...... ·· .. --.-....... ·.···.·· . ·----:-. .--:-:-:-:-:-:-: .. :-:-:-. ·-: .-:-:-.-:-:-:-:-:-:·-:-:-:-:-·-:-:-.-:-
I I GC·FID GOMS EI
.·.· .. -.-.-:-.
I Pro6le Computatioo
The overall experimental procedure for the profiling of primary organics.
58
Table 3.3 Engine sampling series at 1000 rpm
Percentage of Fuel Sampling Water Oil Coolant full load consumption time temperature temperature tempemh1re
.(mllmin) (seconds) ("C) ("C) C'C)
1 9.72 30.20 94 45 4
25 18.93 30.18 95 45 4
50 28.04 30.21 97 so 4
75 38.22 30.32 98 50 4
100 49.18 30.35 941 50 4
1 Fans cut in
Table 3.4 Engine sampling series at 2000 rpm
Percentage of Fuel Sampling Water Oil Coolant full load consumption time tempemture temperature tempemture
. (mllmin) (seconds) ("q ("C) C'q
I 20.62 30.26 92 58 4
25 37.74 30.10 93 58 4
50 58.25 30.40 93 59 4
75 76.92 30.39 93 60 4
lOO 92.31 29.23 95 61 4
Table 3.5 Engine sampling series at 3000 rpm
Percentage of Fuel Sampling Water Oil Coolant full load consumption time temperature temperature temperature
(mi/min) (seconds) ("C) ("C) ("q
1 37.74 29.74 94 58 3
25 65.93 30.12 93 60 3
50 96.77 30.08 94 60 4
75 133.33 30.19 95 60 4
100 176.47 29.92 lOO 72 4
59
3.2.2 Sample Work-up
In order to avoid contamination, all solvents were of HPLC grade (Rathburn Chemicals) and
the solvents were regularly checked for purity by GC-FID. Prior to use, all glassware was
soaked in Decon-90 for 24 hours, rinsed repeatedly in tap water and then distilled water, and
finally dried. The glassware was solvent washed with the appropriate solvents (normally DCM)
before being used in the experiments. All filters and work-up chemicals (silica and drying
agents) were soxhlet extracted with DCM (24 hours) before storing in ovens. All standard
solutions (96+% pure) and samples were stored in freezers (-20°C).
3.2.2. 1 Internal Standards for Aromatics
The initial validation and early experimental use of TESSA by Trier (1988), found the
laboratory recovery of phenanthrene and other three to five ring P AH to be around 90% . The
more volatile two ring naphthalene was recovered by 50%. Therefore, the search for internal
standards for PAH relied on establishing a suitable compound to account for naphthalene and
volatiles, and a more stable compound to account for the remaining more stable P AH.
Deuterated PAH standards were investigated for suitability as internal standards, since the
heavier compound follows the behaviour of the native undeuterated PAH and can be easily
detected by GC/MS El (Levin et al. 1984). The following deuterated standards were
investigated: d8-naphthalene, d10-phenanthrene, and d12-chrysene (Aidrich Chemical Co.).
Following the research of Williams et al. (1986) 9-phenylanthracene and 1 ,3,5-
triphenylbenzene (Aldrich Chemical Co.) were also investigated as internal standards.
To verify the suitability of the internal standards, two solvent mixtures (DCM:methanol,
1: 1 )(800 ml) were spiked with a standard mix containing the EP A 16 priority pollutant PAH
(Supelchem), dibenzothiophene (Janssen Chimica), 9-phenlyanthracene, and 1,3,5-
triphenylbenzene. Distilled water (800 ml) was added to the solvents and the mixture worked
up as forTES. The samples were analyzed by a Varian 6000 GC-FID fitted with a DB-5
capillary column (30 m x 0.32 mm ID, 0.25 ~tm film thickness, J&W Scientific) with the
detector signal recorded and integrated by a Varian 402 data system. Splitless injection at
2400C was used and the temperature program was: 40°C for 1 min., 5°C/min to 300°C, held
for 15 mins. Nitrogen was used for the carrier gas.
60
The quantification was based on response factors to undecane (Janssen Chimica). As shown
in Table 3.6 the majority of the PAH were recovered to the same high level, except
naphthalene and the very high relative molecular mass P AH.
T bl 3 6 a e R I f I esu ts rom ntem al S d d & PAH Lab tanar oratory R ecovery E xpenment
Compound Weight of Laboratory Laboratory Average spike added recovery #1/% recovery #2/% recovery/% (pg)
Naphthalene 200 21.9 14.1 18.0
Acenaphthene 200 72.1 66.8 69.5
Acena_phthylene 200 74.6 7l.S 73.1
Fluorene 200 77.S 77.S 77.S
Dibenzothiojlhene 570 82.0 81.6 81.8
Phenanthrene 200 80.1 82.6 81.4
Anthracene 200 81.3 84.7 83.0
Fluoranthene 200 82.1 85.2 83.7
Pyrene 200 81.0 84.0 82.S
9-Phenylanthracene 210 80.0 83.3 81.7
Benz( u )anthracene 200 82.6 93.6 88.1
Chrayene 200 82.2 86.3 84.3
Benz(i&k)fluoranthene1 200 each 81.0 84.4 82.7
Benzo( u )pyrene 200 78.7 82.2 80.S
1,3,5-Triphenylbenzene liS 76.9 80.3 78.6
Indeno(l,2,3- 200 81.2 85.1 83.2 cd)pyrene
Benzo(ghi)perylene 200 76.4 80.0 78.2
Dibenzo(a,h)anthracene 200 60.6 59.8 60.2
1 Isomers eo-eluted
A selection of different internal standards (Table 3. 7) were added toTES immediately after
collection and to fuel samples, allowing a comparison of the recovery of internal standards
within the sample matrix with the spiked recovery experiments. The results of the review are
given in Section 3.2.2.6.
61
Table 3.7 fi" Internal Standards added to TES & Fuel for Primarv Emission Pro hng
Internal Standard 1000 rpm Engine 2000 rpm Engine 3000 rpm Engine Fuel samples Series Series Series
d.·nanhthalene ./ ./ ./ ./
d,n·Dhenanthrene ./ ./
d .. -chrsvene ./ ./
9·Phenvlanthracene ./ ./ ./
1 ,3,S· Triohenvlbenzene ./ ./ ./
Key: ./ Internal standanl added to sample
3 2 2.2 Extraction and Concentration of TES
The procedures developed by Trier 1988 were used for extraction and concentration of TES.
Firstly, the samples were filtered under reduced pressure (Whatman GF/Fm, 5 cm). The two
phases (DCM and the aqueous methanol) were separated by liquid/liquid partition. The lower
DCM phase was collected in a round bottom flask (RBF, I I) and the remaining aqueous layer
re-extracted with three aliquots of DCM (50 m!).
The DCM fraction was reduced in volume by rotary evaporation to ea. 250 ml before adding
sodium sulphate drying agent (3 g, BDH). After a period of ea. 1·2 hours, the samples were
filtered (Whatman A 11 cm) and reduced by rotary evaporation to leave ea. 0.5 m!. Samples
were transferred to pre·weighed vials and the remaining solvent removed by gently passing a
stream of dry nitrogen over the sample until a constant weight was recorded.
3,2 2 3 Open column Silica gel Fractjonation
The use of gravity fed open columns packed with various stationary phases to fractionate
organic samples into different polarity classes has been widely used (Snook et al. 1979, Lee
et al. 1981, Robbins & McElroy 1982, Schuetzle 1983, Sorrell & Reding 1985, Liberti et al.
1986, and Williams et al. 1989). The method of Trier (1988) was modified. Difficulties in
maintaining reproducibility caused by varying levels of reactive sites have been reported by
Later et al. 1985.
62
To sustain the reproducibility from batch to batch and reduce irreversible adsorption, the silica
(60-120 mesh, BDH) was fully activated by oven drying (24 hours, 185°C), partially
deactivated to optimum activity (5% milli-Q water, w/w) and mechanically shaken for two
hours.
Glass columns (200 mm x 10 mm ID, with glass frits or plugged with defatted cotton wool)
were slurry packed with silica gel (3 g of silica in 20 ml hexane). A plug (ea. 0.5 cm in
height) of sand (acid washed) was applied to the top of the silica bed. The sample (10 to 50
mg) was dissolved in hexane (ea. 0.5 ml) and applied to the column. The vial was rinsed thrice
with hexane (ea. 0.5 ml) and the washings added to the column. The aliphatic class of
compounds were eluted with hexane (12 ml), aromatics eluted with DCM (15 ml), and
remaining polar compounds removed with methanol (25 ml). Before eluting the aromatic and
polar fractions, the sample vial was rinsed with the relevant solvent thrice and washings added
to the column. The different fractions collected in RBFs (50 ml) were rotary evaporated to ea.
0.5 ml, transferred to a pre-weighed vial, and blown down to a constant weight using a gentle
stream of nitrogen. The n-alkanes in the aliphatic fractions and the major PAC in the aromatic
fractions were analyzed by GC-FID (Section 3.2.3) and GC/MS in the El mode (Section
3.2.4) respectively.
3.2.2.4 Fuel work-up
Two fuel samples, collected during the engine sampling period, were worked-up and analyzed
as follows. For each fuel sample, an aliquot of fuel (20 - 40 mg) plus a range of internal
standards (fable 3.7) were fractionated by silica slurry gel adsorption chromatography (Section
3.2.2.3). The aliphatic and aromatic fractions of the fuel were analyzed by GC-FID (Section
3.2.3) and GC/MS El (Section 3.2.4) respectively.
3.2.2.5 Lubricating Oil work-up
An aliquot of fresh oil (66.3 mg) plus deuterated naphthalene (190 J.tg) and deuterated
phenanthrene (170 ~-tg) were fractionated by silica gel adsorption chromatography (Section
3.2.2.3). The aliphatic and aromatic fractions were analyzed by GC-FID (Section 3.2.3) and
GC/MS El (Section 3.2.4) respectively.
63
Before sump oil samples could be fractionated, the particulates in the oil had to be removed.
To allow the particulates to be filtered off, the dispersants first had to be chemically broken
down. This was achieved following the method of Trier (1988). Since three engine series were
collected consecutively, the oil sump sample taken after the last engine series was worked-up
and analyzed (corresponding to 105 hours of engine use).
An quantity of the used sump oil (0.2025 g) was placed in a conical flask. Deuterated
naphthalene (200 l'g) and deuterated phenanthrene (200 l'g) were added. Concentrated
hydrochloric acid (0.5 ml), chloroform (10 ml), and propan-2-ol (10 ml) were added, and the
solution shaken. After five minutes the solution was very slowly filtered under reduced
pressure (Whatman GF/F 5 cm) to remove the coagulated particulates. The filter was washed
with two aliquots of chloroform (5 ml) and excess distilled water added (20 ml). The solution
was then partitioned with hexane (10 ml) to extract the organics. The remaining solution was
re-extracted three times with hexane (10 ml). The hexane solution was dried with sodium
sulphate (3 g) and after a period of time (30 minutes) the drying agent was filtered off
(Whatman A 11 cm). The hexane was reduced by rotary evaporation to ea. 0.5 ml. The
remaining hexane was transferred to a pre-weighed vial and the hexane removed with a gentle
supply of nitrogen, until a constant weight was recorded. A sub-sample (50. 7 mg) of the total
extract was then fractionated by silica gel adsorption chromatography (Section 3.2.2.3). The
aliphatic and aromatic fractions of the sump oil were analyzed by GC-FID (Section 3.2.3) and
GC/MS El (Section 3.2.4) respectively.
3.2.2.6 Recovery of Aromatic Internal Standards in TES
Table 3. 8 shows the levels of the average recovery for the internal standards added to the
engine series and the fuels. The samples were analyzed by GC/MS El and the quantification
was based on the response factors to undecane (Section 3.2.4). The level of recovery were
high, with an average of ea. 86% for all the internal standards used (apart from d8-
naphthalene). The recovery results showed that future sampling could rely on d8-naphthalene
and d10-phenanthrene for laboratory loss correction. The low laboratory recovery of d8-
naphthalene for some samples precluded naphthalene from the initial profiling.
64
Table 3.8 Laboratory Recovery for Internal Standards added toTES and Fuels
Sample Laboratory Laboratory Laboratory Laboratory Laboratory recovery of d8- recovery of d 10- recovery of recovery of 9- recovery of naphthalene phenanthrene d12-chrsyene phenylanthracene 1,3,5-
I% 1% triphenylbenzene 1% 1% 1%
1000 2l.S - 63.4 89.9+ 83.2 qnn
2000 5.4 - !09.0+ - I 14.8 rpm
3000 !0.1 6t.z+ - 71.8 -qnn
Fuels 49.1 84.9 - 94.1 + 89.7
Key : + indicates which internal standards used to correct the PAC quantifications for laboratory losses.
3 2 2 7 I .11horatory Losses for n-Alkanes
The laboratory losses of selected alkanes was investigated by spiking two sets of solvent
mixtures (DCM: methanol, I: I )(800 ml) with a ten times dilution of the n-alkane standard mix
(Section 3.2.3) (250 ~-tl). The solvent mixtures were added to distilled water (800 m!) and the
mixture worked-up as forTES. Following work-up, the samples were diluted with DCM to
250 ~-tl. The areas for the alkanes from before and after work-up were compared using GC-FID
(instrument details in Section 3.2.3). Table 3.9 shows that the laboratory recovery for n
alkanes of molecular weight greater than Cl3 are good.
Table 3.9 Laboratory Recovery of n-Alkanes from Spiking Experiment
Compound Weight Laboratory Laboratory Average I % of spike Recovery from Recovery from added Sample 11% Sample 21%
Undecane 100.0 31.7 49.0 40.4
Tetradecane 150.0 72.0 84.2 78.1
Hexadecane 87.5 86.1 93.5 89.8
Octadecane 175.0 101.8 107.9 104.9
Docosane 162.5 102.7 100.1 100.4
65
3.2.3 Quantification of n-Alkanes by Gas Chromatography with Flame Ionisation Detection
Gas chromatography with flame ionisation detection (GC-FID) is probably the most widely
used GC detection system, due to the large linear range, sensitivity to virtually all organics,
and stable baseline (Lee et al. 1981, Braithwaite & Smith 1985, Uden 1987, and Fetzer 1989).
Following the work-up and isolation of the aliphatic fraction from TES (Section 3.2), the
aliphatic fraction were blow down to a constant weight. The samples (10 mg/ml in DCM) were
analyzed on a Carlo Erba Mega GC fitted with a DB-5 capillary column (30 m x 0.32 mm ID,
0.25 ~-tm film thickness, J&W Scientific). Injection volumes were 0.5 ~-tl and cooled on-column
injection was used. The temperature programme was: sooc (1 min), 5°C/min to 300°C, held
for 15 mins. The detector signal was recorded and integrated by a Shimadzu integrator.
Hydrogen was used as the carrier gas.
The gas chromatograph was calibrated with ann-alkane standard mix (Table 3.10 & Figure
3.3). For calibration purposes the temperature ramp was increased to l0°C/minute and the top
temperature reduced to 270°C without any hold time.
Ta bl 3 10 e n-A lkan d d e stan ar miX U sed fi al . d or an lysis an recovery expenments
Compound Concentration (mg/ml)
Undecane 0.40
Dodecane 0.90
Tridecane 0.35
Tetradecane 0.60
Hexadecane 0.35
Heptadecane 0.85
Octadecane 0.70
Docosane 0.65
This allowed the calibration to be rapidly checked. The quantification was based on
heptadecane (Figure 3.4). The results were expressed as the weight of n-alkane emitted relative
to the amount of that n-alkane in the fuel (Section 4.2.1).
66
2 100 -
4
1 5
3
r-
I lOO
Temperature ( ° C )
6
7
I 200
Key
l = Undecane
2 = Dodecane
3 = Tridecane
4 = T etradecane
5 = Hexadecane
6 = Heptadecan e
7 = Octadecane
8 = Docosane
8
Figure 3.3 Gas Chromatogram of then-alkane standard mix. GC conditions: DB-5 column, 50°C for l min, l0°C/min. upto 270°C.
67
"' Q) .._ <(
~
"' Q)
a...
2500000
2000000
1500000
1000000
500000
0
y = 1 0764.32x-3091.7 R- sq = 0.9988
50 100 150
Weight lniected (ng)
200
Note: Duplicate injections at each weight
Figure 3.4 Calibration curve for heptadecane using gas chromatography with flame ionisation detection.
68
250
3.2.4 Quantification of Major PAC by Gas Chromatography/Mass Spectrometry operated in
the Electron Impact mode
Gas chromatography/mass spectrometry (GC/MS) operated in the electron impact (El) mode
is a powerful analytical tool for PAC analysis (Snook et al. 1979, Lee et al. 1981, Schuetzle
1983, Hites 1989, and Castello & Gerbino 1993). The mass spectrometer detector allows a far
greater confirmation of compounds than GC retention data alone.
The aromatic fractions (5 mg/ml in DCM) from the silica gel clean-up were analyzed for
major PAC using a Carlo Erba 5160 gas chromatograph connected to a Kratos MS25 El mass
spectrometer and managed by a Kratos DS90 computer system. A Supelchem PTE-5 fused
silica capillary column (25 m x 0.32 mm ID, film thickness of 0.25 p.m) was used. The
following chromatographic conditions were applied: 40°C (1 min.), 6°C/min. to 300°C, held
for 15 mins., cooled on-column injection, detector temperature of 300°C, helium carrier gas,
ionizing potential at 40 eV with scanning from 40-568 amu every 2 sec. Injection volumes
were 0.5 p.l.
Under low ionisation potential the major ions for PAC are those of the molecular ion.
Scanning the total ion chromatogram (TIC) for molecular ions within the retention window for
compounds (derived from the standard mix) greatly aids the identification of PAC (Figure
3.5). Different compounds fragment in a unique spectra, providing more confident
identifications than purely retention matches, when the spectra of peaks matches the spectra
of standards (Figure 3.6). Methyl-PAC were also identified on a molecular ion basis (Figure
3.7). The peak areas were manually integrated, with each methyl-PAC separately integrated.
69
100
50
100 ~ u
= 50 ~
-= = = 100 .Q
< ~
50 .~ -~ -~ ~ 100
50
100
50
7.58
Total Ion Chromatogram (TIC):
Molecular ion for pyrene (202) -
Molecular ion for phenanthrene (178)-
Molecular ion for fluorene (166) -
10.38 13.19 15.59 18.40 21.20
Retention Time (minutes)
Molecular ion for dibenzothiophene (184)
24.01 26.41 29.22
Figure 3.5 The identification of PAC on a retention and molecular ion basis . GC conditions: PTE-5 column, 40°C for 1 min., 6°C/min. up to 300°C, held for 15 m ins.
70
202 100
80
60 (a) 101
40
;a.., 20 ·"= 88 Ill 40 57 150 174 = 0 ~ -= 50 100 150 200 250 300 350 .... ~ .::
202 -(If 100 -~ Cl:
80
60 (b) 40
101 20
150 174 0
50 lOO 150 200 250 300 350
Mass/Charge (m/z)
Figure 3.6 The confirmation of pyrene by matching the spectra in the sample (a) to that of the standard spectra stored in the computer I ibrary (b).
71
100
50
100 ~ <.I
= ~ 50 "0
= = .c ~ 100 ~ > ·--~ 50 ~ cc:
100
50
0
- trimethylphenanthrenes (220)
- dimethylphenanthrenes (206)
- monomethy lphenanthrenes ( 192)
- phenanthrene ( 178)
180 205
Temperature ( ° C )
230
Figure 3. 7 Identification of phenanthrene and methylphenanthrenes by the respective molecular ions, 178 = phenanthrene, 192 = monomethylphenanthrenes, 206 =
dimethylphenanthrenes, and trimethylphenanthrenes = 220. GC conditions: PTE-5 column, 40°C for 1 min. , 6°C/min. upto 300°C, held for 15 mins .
72
Quantification was based on response factors to a GC/MS undecane internal standard (stock
concentration of 200 14g/ml) with a quarter of the 5 mg/ml sample dilution derived from the
undecane solution. Internal standard quantification removes the problem of the ionisation
changing for different samples, since the internal standard follows the behaviour of the
compounds of interest (Reece & Scott 1987). The response factors were calculated with a
standard mix containing undecane, 16 EPA PAR priority pollutants, dibenwthiophene, 9-
phenylanthracene, and 1 ,3,5-triphenylbenzene (Table 3.11 and Figure 3.8).
The integration of undecane by the molecular ion was found to be unreliable since undecane
fragments severely, and undecane was quantified using the TIC area, since there was good
resolution from any interfering ions. After correcting for laboratory losses (using the internal
standards identified in Table 3.8) the PAC emissions were expressed relative to the PAC in
the fuel (Section 4.2.1).
Table 3.11 Response factors of PACto Undecane
Compound Response factor #I Resoonse factor #2 Aveme:e Resoonse factor
Naphthalene 0.242 0.206 0.224
Fluorene 0.253 0.207 0.230
Dibenzothiophene 0.319 0.306 0.313
Phenanthrene 0.278 0.249 0.264
Pyrene 0.295 0.256 0.276
73
Key
Undecane 11 = 9-Pheny !anthracene
2 Naphthalene 12 = Benzo( a)anthracene
3 Acenaphthene 13 = Chrsyene
4 Acenaphthy lene 14 = BenzoU&k)fluoranthene
5 Fluorene (not resolved)
6 Dibenzothiophene 15 = Benzo(a)pyrene
7 Phenanthrene 16 = 1 ,3,5-Tripheny Jbenzene
8 Anthracene 17 = Indeno( 1 ,2,3-cd)pyrene
9 Fluoranthene 18 = Benzo(ghi)perylene
10 = Pyre ne 19 = Dibenzo( a, h )anthracene
100 4
Total Ion Chromatogram (TIC) -6
90 - 7 5 13
80 - 3 10 14
70
4J
8 12 16 - 1 15 9
u 60 = - 2 ~ '0
= 50 = -.c < 4J 40
18 19 -
.::: 17 -~ 30 "ii -~
20
11 10 - J~ l \., .._____. \J \
0 I I I
80 160 240
Temperature ( ° C )
Figure 3.8 Total ion chromatogram of the standard PAC mix used for the primary PAC profiling. GC conditions: PTE-5 column, 40°C for L min. , 6°C/min. upto 300°C, held for 15 mins.
74
3.3 Relocation and Modification of TESSA
Following re-location of TESSA due to departmental re-arrangements, a substantial part of the
research was allocated to upgrading the TESSA system. The improvements are documented
in Section 3.3.1 and the initial testing of the new system is reported in Section 3.3.2.
3.3.1 Improvements to the Engine Sampling Rig
One of the first improvements to TESSA was to commission welding of stainless steel flanges
to the three sections of tower; greatly improving the sealing of the tower.
Polytetrafluoroethylene cord (soxhlet extracted for 24 hours with DCM) was placed between
the flanges and the sections tightened with 8 self tightened screws. The flanges enable TESSA
to be easily stripped down and maintained. The flange system enables the bottom section to
be rotated by 180° to allow another engine to be sampled by TESSA. The new system also
allows an extended transfer pipe to be installed between the engine and TESSA, enabling post
combustion reactions to be increased (Section 6.3). The dynamometer was mounted on a track,
allowing a parallel future engine to be controlled.
Whilst TESSA was modified, the glass tubing normally inside the middle section of the tower
were removed and thoroughly cleaned in Decon-90, as were the insides of the tower section.
It was noted that the inside of TESSA and the glass tubing showed little evidence of any
deposits.
One of the drawbacks of the gravity feed solvent system was that in order to maintain a
constant flow rate for each sample, the solvent reservoir had to be filled for each sample. If
this was not the case, the solvent flow would decrease with sampling, and could lead to
differential sampling efficiencies. In an attempt to overcome the gravity feed problem, a
greater head pressure was generated by forcing the solvent mixture from the reservoir using
compressed air.
75
To improve the collection efficiencies a spray head arrangement was designed to finely
disperse the solvent mixture within the oncoming exhaust (Plate 1). The plate positioned close
to the exit of the tube deflects the solvents upwards, creating a fine spray. The shape of the
spray was visualized using water, and the spray adjusted by altering the gap between the end
of the tube and the deflector plate. The spray was set such that the solvents avoided impinging
on the sides of the tower. The solvent flow rate was controlled by a precision flow controller,
and solvent flow rates for varying reservoir pressures and flow controller settings were
established. From these experiments it was found that a pressure of 0.6 Bar and five turns on
the flow-controller would provide a flow rate of 1 1/min.
Since exhaust cooling may increase collection efficiencies (Kruzel et a/1991), a number of
different cooling arrangements were attempted. The first method was to immerse the solvent
reservoir in a mixture of ice and sodium chloride. This mixture produced temperatures as low
as -20"C. However, problems were encountered with regards to achieving an even cooling of
the reservoir, loading/unloading the cooling mixture and the potential for corrosion of the
solvent reservoir. This led to chilling the solvent mix overnight in the freezer (-20°C},
followed by filling the reservoir immediately before sampling. The cooling mixture
(ice:sodium chloride) was used successfully in the chiller tank, decreasing the temperature of
the cooling water around TESSA from 4°C (ice only) to between oac and -5°C.
The upgraded set-up enabled the tower to be wall-mounted closer to the engine exhaust
manifold, allowing a shorter transfer pipe to be fitted (Plate 2). The switching system for
transferring the exhaust from the main exhaust to the TESSA was upgraded to an compressed
air activated device.
A thermometer was installed in-line with the engine air intake. This along with taking the
humidity of the test-cell during sampling sessions allowed the effect of the test cell
environment to be taken into consideration.
The overall engine test bay, showing the position of TESSA relative to the engine, Js
illustrated by Plate 3.
76
mm
Plate 1 The spray head attachment used to evenly disperse the solvents with the oncoming exhaust.
Plate 2 The transfer line from the engine to TESSA.
77
1
Plate 3 The overall engine test bay, showing the upgraded TESSA.
78
3.3.2 Initial Testing of the Upgraded TESSA
The introduction of a pressurized solvent feed system to TESSA was aimed at eliminating the
necessity to fill the solvent reservoir for each sampling position. However, initial testing
showed that the quantity of solvent collected for consecutive runs was not the same. This was
due to pressurization of the solvent mix, causing the DCM component of the solvent mix to
volatilise, and be lost from the solvent supply system. This volatility problem was exaggerated
with the period of time the solvent was exposed to the compressed air. This necessitated a
return to filling the reservoir for each sample. However, the procedure for filling the solvent
reservoir was far simpler than with the initial TESSA system. The initial configuration of
TESSA required raising/lowering the solvent reservoir to a mounting point above the top of
TESSA.
After repeated washing of the tower with solvent mixture, through both solvent inputs, the
upgraded TESSA was evaluated. Initially sampling was at 1000 rpm and at 1%, 8%, 17%,
and 25% of full engine load. The solvent reservoir was filled for each sample with 600 ml of
solvent mix. The solvent flow through the spray head was started 10 seconds before the
exhaust was sampled for 15 seconds using TESSA. The residual solvent was used to wash the
tower down. This initial testing was used to compare the extraction efficiency of the upgraded
TESSA system with the initial TESSA configuration, and to investigate further the effect of
low loads and speeds on the emissions. The samples were worked-up and the aromatic
fractions isolated following the procedures detailed in Section 3.2.2. The aromatic fractions
were analyzed by the Carlo Erba GC-FID (instrument details in Section 3.2.3) and fluorene
was quantified relative to d10-fluorene (98%, British Greyhound), after laboratory loss
correction using d10-phenanthrene (97%, Aldrich Chemical Co.). The recovery of fluorene in
the exhaust relative to the fluorene in the fuel burned, was 0.91% at 1% of full load and
0.26% at 25% of full load for the initial TESSA configuration. The value for the upgraded
TESSA was 0.96% at I% of full load and 0.28% at 25% of full load. This showed that the
new system was working as effectively as the old configuration. The examination of the effect
of low loads at low speed on combustion characteristics, is discussed in the subsequent
chapters.
79
3 .4 Profiling of Secondary Emissions using the upgraded TESSA
The profiling of the secondary emissions was aimed ultimately at the highly mutagenic nitro
PAC, but some preliminary investigations into the emissions of oxy-PAC were performed. The
overall procedure for the secondary emission profiling research is shown in Figure 3.9. The
engine sampling for nitro-PAC, work-up, and analysis are detailed in Sections 3.4.1, 3.4.2
and 3.4.3 respectively.
3.4.1 Engine Sampling for Nitro-PAC Profiling
Sampling for nitro-PAC was proposed to cover a wide range of NO, conditions, in order to
examine the nitration potential of diesel combustion. As can be seen from Figure 3.10 engine
load is strongly correlated with NO., whereas engine speed has a smaller effect. As a
consequence of the large variation in NO, with engine load, the initial sampling for nitro-PAC
was carried out at mid-speed (2500 rpm) and low-, mid-, and high-NO, positions (Section
3.4.1.1). Later engine sampling was undertaken at 1500 rpm and 3500 rpm, and at the same
NO, positions (Section 3.4.1.2).
3.4.1.1 Engine Sampling for Nitro-PAC at 2500 rpm
The tower was cooled using ice and sodium chloride in the chiller. The solvents were cooled
overnight in a freezer (-20°C). The engine was conditioned for one hour at 3000 rpm and 90
Nm. The load positions corresponding to low-, mid-, and high-NO, were calculated from the
NO, engine map (Rhead et al. 1991). The engine was conditioned for 30 minutes at the chosen
NO, position. The longer conditioning time, compared to that used for the primary emission
profiling, was to ensure that the oxides of nitrogen had reached equilibrium (Pipho et al.
1991). The solvent reservoir was fllled with 0.8 L of each solvent, and the solvent flow rate
set to 1 1/min using the reservoir pressure and flow controller. The solvent flow was started
20 seconds before switching the exhaust into TESSA and the exhaust was sampled for 50
seconds. The remaining solvent was used to wash the tower down.
80
Pnma u mplco.l with ·n :s.c;,\
&ir.~.:lion a nc.l Cnnu:utr.~tiun
Clean-up
I
Alipl .. ti<s Aron,.ti<s
GC-AD
Pro fLie Computation
I
PAH. PASH. & ~AH
.-.-.-.-.-·-.-:- ...... -.-··.·.·-·.···-····· • :•: , ,,::• , ,••,•,,,•':,: :: ,•u
GC/"'8 El
NI' III'LC = Norn,.l pl.,~e high perlnnn>nu: liquiU <hr<ln .. ltl!!,r.!(lhy
GC. AD = Ga.• chrou,. togr.~plrywith flame inn;.;,. lio n c.letectiuu
GC-ECD = Gas ch.ruu .. togr.~phy with electron <apture c.lctcctiou
GC/MS El = Gas ~:hrmnatogr.~plty/n,.ss spcctn>mctry opcr.!tL'c.i in elect ron impaclniO<!c
GC/MS NI Cl = GC/MS upcr.~tcc.l iu negative ion d!t!nOOII ionisation mode
I
Polar.;
_l
GC/MSEI
I
ProfLie Computat ion
I J
·.-... -.-.-:--.·.·.·.···· ·-·---·.-.-.-.·.· ... .-.-.·.· .. ·.·: :-.:-.·.···
j GC/MS El
Figure 3. 9 The overall experimental procedure aimed at profiling the secondary nitro- and oxy-PAC.
81
1200
~ 1500 rpm
1000 -·-2500 rpm
-E ...... 3500 rpm c. c. - BOO Cl)
> Cl)
>< 600 0 z ..... en
400 :::J tU
.s= >< w
200
0 ~------~----~----~-----r----~ 0 20 40 60 80 100
Percentage of Full Load
Figure 3.10 The effect of engine speed and load on the gaseous oxides of nitrogen, NOx.
82
Following the addition of d8-naphthalene and d10-phenanthrene internal standards for the
laboratory loss correction of major PAC, the samples were worked-up following the
procedures detailed in Section 3.4.2. For each set of duplicate engine samples, fuel, sump oil
and bosch samples were taken. Full details of sampling at 2500 rpm are given in Table 3.12.
Table 3.12 Engine sampling for nitro-PAC at 2500 rpm
Low- Low- Mid- Mid- High- High-NO, #1 NO, #2 NO, #1 NO, #2 NO, #1 N0,#2
Percentage of full load 1.4 1.4 50 so 85 85
Fue1consurn~ion(niVoUn) 30.73 31.10 74.07 73.17 111.61 110.38
Air intake temperature •c 17 18 16 16 22.5 23.0
Chiller temperature •c -5.0 -4.0 -5.4 -5.2 0.3 0.3
Water temperature •c 96 97 96 95 100 lOO
Oil temperature •c 54 56 52 53 62 62
Bosch No. 1.35 0.98 2.59
Humidity % 62.9 63.1 44.3 45.2 36.5 36.7
Weight of d8-naphthalene 220 220 220 220 540 540 added {p.g)
Weight of d10-pbenanthrene 165 165 165 165 420 420 added {Jlg)
3.4.1.2 Engine Sampling for Nitro-PAC at 1500 rpm & 3500 rpm
The sampling procedure at 1500 rpm and 3500 rpm was the same as used for 2500 rpm , with
the following exceptions. Firstly , fresh oil was used before sampling at 1500 rpm. Secondly,
at 1500 rpm, exhaust samples were taken at low-NOx and mjd-NOx on the same day. The next
sampling session at 3500 rpm also sampled low-NOx and mid-NOx sampling positions on the
same day. Thlrdly, the hlgh-NOx samples at 1500 rpm and 3500 rpm were taken on the same
day. Following the addition of d8-naphthalene and d10-phenanthrene internal standards for
major PAC laboratory loss correction, the samples were worked up as documented in Section
3.4.2. F uel , sump oil, and bosch samples were taken on each sampling day (full sampling
details given in Tables 3.13 & 3.14).
83
Table 3.13 Engine sampling for nitro-PAC at 1500 rpm
Low-t Low-t Mid-t Mid-t High- High-NO #1 NO #2 NO #1 NO, #2 NO #1 NO #2
Percentage of full load 1.8 1.8 50 50 85.6 85.6
Fuel consumption (mVmin) 16.42 16.68 39.28 38.94 62.1 1 62.42
Air intake temperature oc 22 23 22 23 24 23
Chiller temperature °C -1.1 2.4 4 3.9 -0.5 -0.1
Water temperature oc 95 95 98 97 100 98
Oil temperature °C 48 47 50 52 50 5 1
Bosch No. 0.93 1.24 3.27
Humidity% 33.4 33.5 32.2 31.4 50.1 51.5
Weight of d8-naphthalene 110 110 270 270 460 460 added (JJ.g)
Weight of d 10-phenanthrene 112 112 210 210 303 303 added (JJ.g)
+collected on same day
Table 3. 14 Engine sampling for nitro-PAC at 3500 rpm
Low-* Low-* Mid-* Mid-t High-* High-* NO #1 NO #2 NO #1 NO #2 NO, #1 NO. #2
Percentage of full load 1.6 1.6 29 29 87 87
Fuel consumption (mVmin) 49.78 50.86 85.11 86.51 172.21 173.31
Air intake temperature °C 23 24 24 24 28 27
Chiller temperature oc -1.0 -1.0 -0.8 -0 .9 1.3 1.8
Water temperature °C 100 100 100 100 110 115
Oil temperature oc 72 70 70 70 90 92
Bosch No. 2.70 2.06 5.75
Humidity % 33.5 32.5 31.4 31.2 50.0 49.0
Weight of d8-naphthalene 416 416 624 624 819 819 added (JJ.g)
Weight of d10-phenanthrene 418 418 543 543 707 707 added (p.g)
tcollected on same day for that speed *Collected on same day for that speed *Collected on same day as the high-NO, samples at 1500 rpm
84
3.4 2 Sample Work-up for Nitro-PAC
The sample work-up for the secondary emission samples followed all the preparation measures
employed for the primary emission profiling. Due to the nitro-PAC being photo sensitive
(Stark et al. 1985), all work-up was carried out under red light, or alternatively, all apparatus
wrapped in foil. All nitro- and oxy-PAC standards were obtained from Aldrich Chemical Co.
and were greater than 97% in purity.
The work-up procedure for exhaust samples consisted of a modified extraction and
concentration procedure to reduce the possibility of nitrating the samples (Section 3.4.2.1).
The samples taken at 2500 rpm were subjected to a simple cartridge clean-up to remove the
aliphatics; whereas the silica gel clean-up for the samples at 1500 rpm and 3500 rpm removed
the aliphatics and polars (Section 3.4.2.2). The aromatic fractions containing the nitro-PAC
were further fractionated using normal-phase HPLC to isolate specific P AC classes (Section
3.4.2.3). The work-up procedures for fuel and oil samples are given in Section 3.4.2.4. The
estimation of laboratory losses for nitro-PAC during the exhaust work-up procedure is detailed
in Section 3.4.2.5.
3 4 2.1 Extraction and Concentration of Exhaust Samples for Nitro-PAC Profiling
The potential of the aqueous methanol phase to nitrate the organics presence in the DCM
fraction was considered a possibility, since the aqueous phase may contain nitric acid, a
powerful nitration agent (Nielsen 1984). With the aim of reducing the nitration possibility, the
samples were partitioned first and filtered second. This was the opposite way to the profiling
of the primary organics. The reversal of the steps meant that the organics could be rapidly
removed from the aqueous phase. The concentration and drying procedures were as for the
primary organics work-up (Section 3.2.2.2).
The nitration potential of the methanol phase was checked for engine samples taken at 1500
rpm and 3500 rpm. After completing removing the DCM fractions, an aliquot (20 - 40 ml) of
the aqueous methanol layer was removed for pH testing. The pH varied between 3.15 and 4.13
for samples.
85
The remaining aqueous methanol was spiked with the standard PAH mix (lOO ~1) used for
GC/MS analysis of major PAH and PASH (Section 3.2.4). The solution was vigorously
shaken and left for 24 hours. The aqueous methanol fraction was then extracted three times
with DCM (20 ml). The solvent was dried with sodium sulphate (1 g, BDH), filtered
(Whatman A 11 cm), and reduced by rotary evaporation to ea. 0.5 ml. After transferring to
a pre-weighed vial the remaining solvent was removed with a gentle flow of nitrogen. The
total extracts were cleaned-up using glass columns packed with silica gel (Section 3.4.2.2.2)
and the mononitro-PAC fraction isolated by normal-phase HPLC (Section 3.4.2.3). After
removing the solvent from the HPLC fractionation and diluting with acetonitrile (50 ~1), the
nitration check was examined by GC-ECD (refer to Section 3.4.3.1.4 for full details of GC
ECD). The presence of nitro-PAC was checked by eo-injection of the nitration check with the
mononitro-PAC standard mix. The results were compared with the solvent blank (Figures 3.11
& 3.12). There was no evidence for any nitro-PAC being formed from the aqueous methanol
fraction.
3.4.2.2 Clean-up of TES
To aid the detection of the trace nitro-PAC, a greater isolation and concentration step of the
nitro-PAC was required than was needed for the detection of major aromatics. The desired
fractionation scheme was for a simple cleanup step to remove the aliphatics, and then apply
the aromatics to HPLC fractionation to isolate the nitro-PAC fractions of interest. Initially a
simple cartridge clean-up step was used (Section 3.4.2.2.1). However, interference problems
associated with aliphatic compounds were later found for the analysis of the mononitro-PAC
HPLC fractions by GC-ECD. This led to reverting back to the silica gel glass column clean-up
method with modified solvent elution volumes (Section 3.4.2.2.2).
3.4 2 2.1 Octadecylsilane Cartridge Clean-up
One method of achieving the desired clean-up was previously used by Obuchi et al. (1984).
This method used octadecylsilane (ODS) cartridges to retain the aliphatic components of crude
oils, whilst eluting the aromatics. This method was tested using ODS Sep-Pak cartridges
(Millipore, sorbent mass of 360 mg) attached to a 20 ml glass syringe and using diesel fuel as
a test sample. Diesel fuel (48.7 mg) was applied to the column and the vial washed with
acetonitrile (0.5 ml), and the washings applied to the cartridge.
86
Key 1 = 1-Nitronapbthalene 2 = 2-metbyl-1-nitronaphtbalene 3 = 2-Nitronapbthalene
21 22
2
23
(a) Nitration Check
(b) Solvent Blank
(c) Nitro-PAC Standards
24 25
Retention Time (minutes)
Figure 3.11 The evaluation of the acidic aqueous methanol phase for the formation of nitronaphthalenes. The comparison of the nitration check phase (a) to both the solvent blank (b) and the mononitro-PAC standard mix (c) proved there was no such nitration occurring. (Chromatographic conditions given in Section 3.4.3.1.4.1).
87
Key I = 3-Nitrofluoranthene 2 = 1-Nitropyrene
1--"'-'------~
2
41 42
(a) Nitration Check
(b) Solvent Blank
-----~-~- ----------
(c) Nitro-PAC Standards
43 44
Rentention Time (minutes)
45
Figure 3.12 The evaluation of the acidic aqueous methanol phase for the formation of 3-nitrofluoranthene and 1-nitropyrene. The comparison of the nitration check (a) to both the solvent blank (b) and the mononitro-PAC standard mix (c) proved there was no such nitration occurring. (Chromatographic conditions given in Section 3.4.3.1.4.1).
88
The cartridge was slowly eluted with 0.5 ml increments of acetonitrile and the fractions
collected. The results shown that as little as 1 to 2 ml of acetonitrile was sufficient to eo-elute
the aliphatics with aromatics.
However, when a TES was used a far greater separation was achieved after 5 - 6 ml of
acetonitrile. A similar effect was found by Obuchi et at (1984). The eventual procedure for
the ODS clean-up was to apply the exhaust sample (ea. 25 - 48 mg) dissolved in acetonitrile
to a pre-rinsed cartridge (20 m! of hexane, followed by 20 m! of DCM, and finally 20 ml of
acetonitrile). The sample was then pushed onto the cartridge and the aromatics eluted with
acetonitrile (6 ml). There was some breakthrough of some lower molecular weight n-alkanes.
It was envisaged that the small breakthrough of aliphatics would be removed with the major
PAH fraction from the HPLC fractionation, and the small aliphatic eo-elution would not hinder
the identification of major aromatics by GC/MS.
However, the analysis of one from each of the duplicate HPLC mononitro-PAC fractions from
the 2500 rpm sampling, by GC-ECD (refer to Sections 3.4.2.3 and 3.4.3.1.4 for details of
HPLC fractionation and GC-ECD respectively) showed, in some cases, the presence of
aliphatics. The negative spiking of the baseline, caused by aliphatics quenching the ECD, upset
the baseline integration markers. The presence of trace aliphatics in the mononitro-PAC
fractions from 2500 rpm was confirmed by GC-FID analysis.
The eo-elution of aliphatics with the aromatic fraction of TES and the subsequent GC-ECD
baseline disturbances led to increasing the hexane segment of the solvent HPLC fractionation
programme (Section 3.4.2.3). This increased the likelihood of eluting the aliphatics in the
major PAH HPLC fraction and not the later mononitro-PAC HPLC fraction. However, the
analysis of the remaining 2500 rpm samples by GC-ECD also found aliphatic interferences in
the mononitro-PAC HPLC fractions.
3.4 2 2 2 Silica gel Clean-up
The aliphatic eo-elution problem associated with the ODS cartridges, led to the samples from
1500 rpm and 3500 rpm being subjected to gravity fed open column silica gel adsorption clean
up (Section 3.2.2.3). The volumes of hexane used for the removal of the aliphatics was
89
increased to 25 ml, and the volume of DCM used to remove the nitro-PAC was increased to
35 ml. This clean-up step also helped to reduce the accumulation of polar material on the
HPLC fractionation column.
3.4.2 3 High Performance Liquid Chromatography Fractionation
Normal phase (NP) high performance liquid chromatography (HPLC) employing different
stationary phases and solvents has been used extensively for isolation of nitro-PAC from diesel
extracts (Levine & Skewes 1982, Paputa-Peck et al. 1983, Schuetzle et al. 1982, Nielsen
1983, Tomkins et al. 1984, Tong et al. 1984, and Draper 1986). Out of all the commercially
available stationary phases, the most widely used is that of silica. HPLC fractionation with a
silica stationary phase allows the retention windows for chemical classes of differing polarity
to be established by standards. One of the problems with silica columns is the irreversible
adsorption of material and water, producing inconsistent retention times. To counter this a
guard column was used and the solvents were stored with calcium hydride (3 g) (40 mesh,
95 +%, Aldrich Chemical Co.). The solvents were filtered prior to use (Whatman 0.5 I'm).
A Perkin Elmer series 410 LC pump connected to a Perkin Elmer LC-235 diode array UV
detector was used, with the data recorded on a Perkin Elmer GP-100 chart recorder. The
detector signal was also transferred to an IBM PC using a PE Nelson 900 series interface and
processed by PE Nelson associated software. The monitoring wavelength was set at 254 nm.
A 10 11m semi-preparative Spherisorb silica (HPLC Technology) guard column (5 cm x 10 mm
ID) and fractionation column (25 cm x 10 mm ID), using the same stationary material, were used
with an 100 111 sample loop fitted to a Rheodyne 7125 injector.
There were two solvent programmes used in this study. The initial solvent programme was used
for the fractionation of the samples collected at 2500 rpm (Section 3.4.2. 3 .I). The initial solvent
programme was modified in an attempt to eliminate aliphatic co-elutions. The modified
programme also separated the dinitro- and nitro-oxy-P AC compounds. The ffactionation of the
aromatic fraction of exhaust samples from 1500 & 3500 rpm by the revised HPLC method is
detailed in Section 3.4.2.3.2.
90
3.4.2.3. 1 Initial HPLC Fractionations and Method Development
The initial solvent programme was aimed at isolating the mononitro-PAC from the
aromatic/polar fraction eluted from the ODS cartridges. The gradient elution programme was
based on the work of Tong et al. (1984). It consisted of 100% hexane for 5 min; programmed
to 100% DCM over 20 min, held for 10 min , programmed to 100% acetonitrile over 10 min,
held for 10 min; programmed back to 100% DCM in 5 min and finally to 100% hexane in
further 5 min. The flow-rate was 4 ml/min. Standard mixes (Table 3.15) were used to establish
the retention windows for the fractions to be collected (Figure 3.13), with the retention times
being reproducible (Table 3.16).
Table 3.15 HPLC fractionation standard mix
Mix Compounds
PAH Naphthalene, fluorene, & phenanthrene
PANH Quinoline, indole, I ,4-napbthaquinone, & carbazole
Mononitro-PAC 1-N itronaphthalene, 2-nitronaphthalene, 2-methyl-1-nitronaphthalene, 2-nitrofluorene, 9-nitroanthracene, 3-nitrofluoranthene, & 1-nitropyrene
Dinitro-PAC 1 ,5-Dinitronaphtha1ene, I ,8-dinitronaphthalene, 2, 7-dinitronaphthalene, & 1,6-dinitropyrene
Nitro-oxy-PAC 3-nitrofluoren-9-one, 2, 7 -dinitrofluoren-9-one, & 2,5, 7 -trinitrofluoren-9-one
Table 3. 16 Reproducibility of Retention Times (R,) for NP HPLC Fractionation
Injection# R. for 1-nitronaphthalene R. for 1,8 dinitro- R. for 3-nitro-fluoren-9-one (minutes) naphthalene (minutes) (minutes)
1 17.31 25 .07 31.40
2 17.31 25.61 31.56
3 17.64 26.00 31.40
4 17.59 25.89 31.24
5 17.26 25.34 31.40
6 17.64 25.95 31.35
Average 17.46 25.64 31.39 (minutes)
Standard 0.167 0.343 0.0942 Deviation
91
e = ..,.
lfi N
Key A = P AH standards B = Mononitro-PAC standards C = Dinitro-PAC & Nitro-oxy-PAC standards
A
10
\
\ \
\ \ \
\ \ ~
I \ \ \ l ~ ,s, \1
20 30
Retention Time (minutes)
c
40
Figure 3.13 The chromatogram of the HPLC fractionation standard mix and the establishment of the initial HPLC fraction elution windows. HPLC conditions: Spherisorb silica column, 100 % hexane (5 min. ), increased to 100% DCM (20 min.) , held 10 min., increased to 100% acetonitrile (10 min) , and held for 10 min. Flowrate of 4ml/min.
92
The aromatic fractions from the ODS cartridge clean-up of one from each duplicate obtained at
2500 rpm were injected (100 J.ll, 100 mglml). Initially four fractions were collected in RBFs (lOO
ml) (Figure 3.14). Preliminary analysis of the HPLC fractionations by GC-ECD, indicated that a
small proportion of aliphatics had broken through on the ODS cartridges. As previously explained
(Section 3.4.2.2.1), the aliphatics had the effect of producing negative peaks on the GC-ECD. In
an attempt to eliminate the aliphatic interferences the hexane part of the solvent programme was
increased to 15 minutes. Preliminary analysis of the third HPLC fraction (which may have
contained dinitro-PAC, nitro-oxy-PAC, and some oxy-PAC) showed the fractions to contain
many compounds. Hence, fraction three of the initial solvent programme was further fractionated
in an attempt to concentrate further the dinitro-PAC. There were now five HPLC fractions. The
modified procedure failed to elute the aliphatics prior to the mononitro-PAC.
3 4 2 3 2 HPLC fractionations of TES collected at 1500 rpm & 3 500 rpm
To remove the aliphatic eo-elution problem the TES collected at 1500 rpm and 3500 rpm were
cleaned-up using glass columns packed with silica gel (Section 3.4.2.2.2.). The aromatic fractions
from the clean-up were further fractionated by NP HPLC using the revised solvent programme.
There was no problem with aliphatic carry over for these samples and five fractions were collected
in RBFs (50 ml and 100 ml).
3 4 2 4 Fuel and Oil Work-up
Following silica gel fractionation (Section 3.4.2.2.2) one fuel sample was subjected to NP HPLC
fractionation, using the revised procedure, for investigation ofnitro-PAC in the fuel. Following
the removal of particulates (Section 3.2.2.5) and the silica gel clean-up (Section 3.4.2.2.2) of
sump oil (obtained after last sample for nitro- & oxy-PAC profiling was taken, and corresponding
to 50 hours of use), the aromatic fraction was subjected to NP HPLC fractionation, using the
revised procedure, to examine the nitro-PAC content of the oil.
93
100
Key
A = P AH, PASH, and alkyl-derivatives B = Monooitro-PAC C = Dioitro-PAC & Nitro-oxy-PAC D =Polar PAC
10 20 30 40
Retention Time (minutes)
50 60
Figure 3.14 The HPLC fractionation of the aromatic fraction from the ODS clean-up of TES collected at 2500 rpm and low load. HPLC conditions: Spherisorb column, 100 % hexane (5 min.) , increased to 100% DCM (20 min.), held lO min., increased to 100 % acetonitrile (10 min), and held for 10 min. Flowrate of 4 mllmin.
94
3.4.2.5 Recovery ofNitro-PAC for Laboratory Work-Up
The recovery ofnitro-PAC for the entire work-up procedure was divided into three stages, with
each stage being carried out in duplicate. Firstly, the HPLC fractionation mix (50 J.!l) was added
to the solvent mixture (DCM:methanol, 1:1, 800 ml total) plus distilled water (800 ml). The
mixture was worked-up to the total extract stage. Similarly in the second stage, the HPLC mix
(50 J.ll) was applied to silica gel columns or ODS cartridges and the samples worked-up. In the
last stage, the HPLC mix was injected (10 J •. tl) on the NP-HPLC system and the different fractions
worked-up. At each stage, the recoveries (Table 3. 17) were calculated by comparing areas of
compounds in the recovery standard mix before and after work-up by the Carlo Erba GC-FID
(instrument details in Section 3.2.3, with nitrogen instead of hydrogen used for the carrier gas).
Table 3.17 Laboratory recovery of nitro-PAC
Compound Laboratory Laboratory Laboratory Laboratory Laboratory recovery recovery recovery recovery recovery for after total after ODS after silica after HPLC total extract clean-up gel clean-up fractional ion procedure work-up (using silica
% % % % clean-up) %
1-nitronaphthalene 95.7 97.6 101.6 92.8 90.1
2-nitronaphthalene 94.6 98.1 101.2 94.5 90.3
2-nitrofluorene 95.7 99.2 100.7 94.2 90.6
1-nitropyrene 97.0 92.6 101 .2 79.2 77.4
I ,8-dinitronaphthalene 97.4 99.9 99. 1 72.2 68.7
2, 7 -dinitronaphthalene 96.8 94.2 93.0 70.4 60.2
3-nitro-fluoren-9-one 102.8 96.0 88.5 84.3 75.6
3 .4.3 Analysis ofNitro- and Oxy-PAC
The analysis ofnitro-PAC explored several detection systems and established GC-ECD as the best
system, with GC/MS NICI used as a back-up technique (Section 3.4.3. 1). The oxy-PAC were
analyzed by GC/MS El (Section 3.4.3.2). The n-alkanes and the major PAH were analyzed by
GC-FID (Seection 3.4.3.3) and GC/MS El respectively (Section 3.4.3.4).
95
3.4.3.1 Evaluation and establishment of detection systems for Nitro-PAC analysis
The following detection systems were evaluated with a mononitro-PAC standard mix (fable
3.18) and in some cases with samples for the analysis of nitro-PAC:
1) Gas chromatography with nitrogen-phosphorous detection (GC-NPD) (Section 3.4.3.1.1)
2) Reverse phase HPLC with fluorescence/chemiluminescence detection (Section 3.4.3. 1.2)
3) Gas chromatography (GC/MS) operated in electron impact (El) mode (Section 3.4.3.1.3)
4) Gas chromatography with electron capture detection (GC-ECD) (Section 3.4.3.1.4)
5) GC/MS operated in the negative ion chemical ionisation (NICD mode (Section 3.4 .3. 1.5)
Table 3.18 Mononitro-PAC standard mix
Compound Stock Concentration (m_g/ml)
1-nitronaphthalene 0.10
2-nitro-1-methyl-naphthalene 0.10
2-nitronaphthaJene 0.10
6-nitroquinoline 0.10
2-nitrotluorene 0.10
9-nitroantbracene 0. 16
3-nitrotluoranthene 0. 13
1-nitropyrene 0.10
3.4.3.1.1 Gas Chromatography with Nitrogen-Phosphorous Detection
Gas chromatography with nitrogen-phosphorous detection (GC-NPD) is a selective detector
for nitrogen and phosphorous compounds, and has been employed for the analysis of nitro
PAC in diesel extracts and environmental samples (Oehme et al. 1982, Nielsen 1983, Paputa
Peck et al. 1983 and Campbell & Lee 1984). The technique is reviewed by White et al.
(1984). The evaluation was performed on a Carlo Erba Mega series GC with a NPD-40
detector fitted . A Supelchem PTE-5 (25 m x 0.32 ID, 0.25 J.Lm film thickness) capillary
column was used with cooled on-column injector. The output of the gas chromatograph was
recorded by a Shimadzu integrator. After trying several different bead heights the sensitivity
was low and the chromatography unstable; leading to GC-NPD being dismissed for nitro-PAC
analysis.
96
3 4 3 1 2 Reverse Phase 1-IPLC with Fluorescence/Chemiluminescence Detection
Reverse phase HPLC with fluorescence/chemiluminescence detection is a sensitive technique for
nitro-PAC, and in the case of chemiluminescence very selective as well. As a consequence of the
greater selectivity, the chemiluminescence technique has been gaining popularity in recent years
(Sigvardson & Birks 1984, Imaizuimi et al. 1990, Liu & Robbat 1991, and Li & Westerholm
1994). Chemiluminescence detection is reviewed by Robards & Worsfold ( 1992) and K wakman
& Brinkman (1992).
For evaluation of fluorescence, a Merck Hitachi system equipped with a L-6000 two pump system
connected to an F-1050 fluorescence detector and linked to a D-2500 integrator was used. A PE
Nelson data computer system was also used to collect and process the data. Sample injection were
made using a Rheodyne 7125 injector with a 20 Ill loop. A Spherisorb-2 ODS 5 1-1m (HPLC
Technology) column (25 cm x 4.6 mm ID) was used. The evaluation was performed using 1-
aminopyrene and 1-nitropyrene. The method of Liu & Robbat (1991) was used, with the
excitation (exl) and emission (eml) wavelengths were 260 nrn and 430 nm respectively.
The detection limits for nitro-PAC can be significantly increased if the nitro group is reduced to
the amino derivative (Sigvardson & Birks 1984, MacCrehan et al. 1988, and Liu & Robbat 1991 ).
This can be achieved on or off-line. In this research the on-line reduction adopted method ofLiu
& Robbat (1991) was used. The reductive columns were made by packing stainless steel columns
(4 cm x 2 mm id) with zinc (99.998%, Aldrich Chemical Co.). In order for the reduction
mechanism to proceed an ammonium acetate buffer was used. Isocratic elution with acetonitrile
water (80:20, v/v) with ammonium acetate buffer (30 mM, 99%, Aldrich Chemical Co.) was used
to evaluate the system. Figure 3. 1 5 shows the high sensitivity of the system for detection of 1-
nitropyrene in the reduced amino- form. The peak shape was acceptable, indicating that the
reductive column was not causing excessive peak broadening.
97
.~ - 1-Arninopyrene
= Qi -~ Qi u
= Qi u <11 Qi loo
= = ~
4 8
Retention Time (minutes)
Figure 3.15 The sensitivity of fluorescence detection with on-line pre-column reduction for the determination of 1-nitropyrene (2.5 pg injected) in the reduced from of 1-aminopyrene. HPLC conditions: Spherisorb ODS-2 column, isocratic elution with acetonitrile:water (80:20) , and ammonium acetate buffer (30 mM). Exft. = 260 nm and Emft. =430 nm. Flowrate of l ml/min.
98
The evaluation of chemiluminescence detection was performed in conjunction with the
development and testing of a purpose built chemiluminescence detector (CamSpec) by
Gachanja (1994), at the University of Plymouth. The chemiluminescence reaction involved
with bis (2,4-dinitrophenyl)oxalate (DNPO) (1.45 mg/ml, in acetonitrile) and hydrogen
peroxide (40% HP2 in acetonitrile, v/v) was initially tested. Flow injection analysis was used,
with the two reagent streams (controlled by periostatic pumps) mixed prior to meeting the
eluent flow. Flow rates of 0.5 ml/min for each reagent stream was used, and the flow rate for
the mobile phase (100% acetonitrile) was 1 ml/min. The sensitivity of the system was
investigated with 1-aminopyrene (Figure 3.16). However, when the chemiluminescence
reagents were combined with the ammonium acetate buffer used for the on-line reduction of
the nitro-PAC, the chemiluminescence reaction did not work. This was a consequence of the
buffer pH of 7.4 inhibiting the chemiluminescence. The incompatibility of the DNPO reaction
with the on-line reduction of nitro-PAC, led to investigating bis (2,4,6- trichlorophenyl)oxalate
(TCPO) as an alternative chemiluminescence reagent. The TCPO reaction is less rapid than
DNPO but has the advantage of optimum sensitivity at a pH of around 7.0 (Imaizumi et al.
1989, Robards & Worsfold 1992, and Gachanja 1994). The same chromatographic conditions
as used for DNPO were employed with the TCPO appraisal. The TCPO system gave similar
sensitivities to 1-aminopyrene as achieved with DNPO and the system was compatible with the
on-reductive HPLC system.
On a routine basis, the analysis suffered greatly from the short lifetime of the zinc reductive
columns, resulting in unacceptable peak broadening. This resulted in noTES being analyzed.
Future reliability could be improved by using a heated rhodium/platinum reduction column
(Tejada et a/.1982 & 1986).
99
0 ·;;; c ~ -c ~
~ C.l c ~ C.l
~ .5 e = -·-e ~
.::: u
2 3 4 5 6
Retention Time (minutes)
Figure 3.16 The sensitivity of the bis (2,4-dinitrophenyl)oxalate (DNPO) chemiluminescence reaction to the detection of 1-aminopyrene (20 ng injected each time) using flow-injection analysis. Flow rate of 1 mllmin for mobile phase (100% acetonitrile) and 0.5 mllmin for each reagent stream (DNPO and H20 2).
100
nitro-P AC and has been used to confinn prominent nitro-P AC in some diesel studies (Schuetzle
et al. I983 and Bayona et al. 1988).
The evaluation of GC/MS in El mode was perfonned on a Carlo Erba 5160 GC linked to a Kratos
MS25 mass spectrometer. The ionisation was set at 40 eV, and 40-528 amu were scanned every
2 seconds. A Rt,.-5 (25 m x 0.32 ID, 0.25 Jlm film thickness) capillary column was used with
cooled on-column injector. The temperature program consisted of 60"C for I minute, 5"C/rnin
temperature ramp up to 300"C, held for 10 minutes. Helium was used for the carrier gas. The
output of the mass spectrometer was managed by a Kratos DS90 data system. The detection
limits were of the order of 0.5 - 1.0 ng per injection for some nitro-PAC, such as 1-
nitronaphthalene (signal/noise, SIN, 3:1) (Figure 3 .17). Injection of mononitro-PAC HPLC
fractions could not detect any nitro-PAC.
3 4 3 I 4 Gas Chromatography with Electron Ca.pture Detection
Gas chromatography with electron capture detection (GC-ECD) is very sensitive towards
electrophilic compounds and has been used for the analysis of nitro-PAC (Oehme et al. 1982,
Campbell & Lee 1984, and Draper et al. 1986). The technique is reviewed by Poole (1982).
The evaluation of GC-ECD was perfonned on a Perkin Elmer 8600 GC with ECD fitted. A
Supelchem PTE-5 (25 m x 0.32 ID, 0.25 Jlm film thickness) capillary column was used with a
programmable temperature (PTV) injector. The output of the gas chromatograph was recorded
by a GP-1 00 Perkin Elmer chart recorder. The output was also linked to an mM compatible PC
by a Perkin Elmer Series 900 data capture unit before sending to the PE Nelson integration
package. The computer system enabled a backup of runs to be made and allowed reprocessing
of data at a later date.
The oven temperature and other parameters were initially based on the work of Oehme et al.
(1982): I minute at 40"C, 30"C/rnin to I40"C, 5"C/min to 300"C, detector temperature 320"C,
injection volumes ofO.S Jll, nitrogen as the carrier gas. The initial analysis was severely hindered
by the detector cell rapidly becoming contaminated.
101
100
~ (j
= ~ "'0
= = .c < ~ > ·--~ ~ ~
Key 1 = 1.0 ng of 1-Nitronaphthalene 2 = 1.0 ng of 2-Methyl-1-nitronaphthalene 3 = 1.0 ng of 2-Nitronaphtbalene 4 = 1.0 ng of 6-Nitroquinoline 5 = 1.0 ng of 2-Nitrofluorene 6 = 1.6 ng of 9-Nitroanthracene 7 = 1.3 ng of 3-Nitrofluorantbene 8 = 1.0 ng of 1-Nitropyrene
2
Total Ion Chromatogram (TIC): 6
5
3
4
3.38 11 .00 18.22 25.44 33.06
Retention Time (minutes)
Figure 3. 17 The low sensitivity of gas chromatography/mass spectrometry operated in the electron impact mode to the mononitro-PAC standard mix. GC conditions: Rlx-5 column, 60°C for I min. , 5°C/min to 300°C, held for 10 min.
102
The contamination affected the saturation levels and consequently the sensitivity. The
contamination problem was greatly reduced by adding a piece of deactivated silica tubing (1
m x 0.32 mm ID, Jones Chromatography) to the detector end of the analytical column. The
columns were connected using a glass column connector (Jones Chromatography). The
deactivated silica column enabled the temperature of the detector to be increased to 350°C,
without significant increase in column bleed. The higher detector temperature reduced the
accumulation of material and thus a lower (and hence more sensitive) saturation level could
be maintained.
The detection limits were of the order of ea. 50 pg per mononitro-PAC injected (SIN of 3/1)
(Figure 3.18). The linear ranged from the detection limits to ea. 2 ng, The reproducibility of
both the retention times and integrations was good. For example, five replicate injections of
1-nitropyrene (50 pg) gave a standard deviation of 0.0964 for the retention times and 0.549
for the integrations respectively. The excellent GC resolution, high sensitivity, and access to
instrument resulted in GC-ECD being used extensively for the detection of nitro-PAC in the
HPLC fractions. The GC-ECD analysis is divided into the mononitro-PAC HPLC fractions
(Section 3.4.3.1.4.1), dinitro-PAC HPLC fractions (Section 3.4.3.1.4.2), and nitro-oxy-PAC
HPLC fractions (3.4.3.1.4.3).
3 4.3.1 4 I Analysis of Mononitro-PAC HPLC Fractions by GC-ECD
This section describes the analysis of the mononitro-PAC HPLC fractions by GC-ECD. The
final temperature programme was 60°C for 1 min., 5°C/min. to 300°C, held for 10 mins ..
Nitrogen was used as the carrier gas. The PTV injector was programmed as follows: 0.01
minutes after injection temperature increased from under 50°C to 300°C, 2.49 minutes later
purge vent opened, and 1.51 minutes later the PTV injector was cooled back to under 50°C
using compressed air.
Integration of peaks within the retention windows established by the mononitro-PAC standard
mix (Figure 3.19) were between 20-25 minutes for nitronaphthalenes, 30-35 minutes for
nitroanthracenes and nitrofluorenes, and 40-45 minutes for nitropyrenes and
nitrofluoranthenes. These windows were used to mark the baseline start and stop positions. In
doing this, the baseline was forced to follow directly underneath the peaks of interest.
103
Key 1 = 50 pg of 1-Nitronaphthalene
2 = 50 pg of 2-Methyl-1-nitronaphthalene
3 = 50 pg of 2-Nitronaphthalene
4 = 65 pg of 3-Nitrofluoranthene
5 = 50 pg of 1-Nitropyrene
4
258
1
160
0 Temperature ( C)
2
172
5
268
Figure 3.18 The high sensitivity of gas chromatography with electron capture detection to a) the nitronaphthalenes, and b) 3-nitrofluoranthene and 1-nitropyrene. GC conditions: Supelchem PTE-5 column, 60°C for 1 min., 5°C/min to 300°C, held for lO min.
104
'T I
I I
10 20
2
6
5
4
I
30
Key l = 1-Nitronaphthalene 2 = 2-Methyl-1-nitronaphthalene 3 = 2-Nitronaphthalene 4 = 6-Nitroquinoline 5 = 2-Nitrofluorene 6 = 9-Nitroanthracene 7 = 3-Nitrofluoranthene 8 = l=Nitropyrene
x = Septum bleed
7
8
X
I
40 5o I
60
Retention Time (minutes}
Figure 3.19 The establishment of the integration temperature windows for nitro-PAC using the mononitro-PAC standard mix. GC conditions: Supelchem PTE-5 column, 60°C for 1 min.' 5°C/min to 300°C, held for 10 min.
105
Mononitro-P AC HPLC fractions were blown down to constant weight under a gentle stream of
nitrogen. The weights were low for all samples; less than 0.2 mg. Sample dilutions were made
with I 00 Jll of acetonitrile. Mononitro-PAC were identified by eo-injecting the sample with the
mononitro-P AC standard mix. The increase of peak areas with respect to the sample injection
only, confirmed the presence of compounds. eo-injections were performed with different
concentrations of the standard mix. In this way the possibility of the standard mix masking several
peaks was eliminated as the concentration of the mix was progressively reduced. Figure 3.20
shows the confirmation of the nitronaphthalenes by eo-injection techniques. Positive eo-injection
confirmations were achieved for 1-nitronaphthalene, 2-nitronaphthalene, 2-methyl-1-
nitronaphthalene, and 1-nitropyrene for all samples. The confirmation of the presence of nitro
p AC was backed-up by retention indices based on 1-nitropyrene (Table 3 .19).
Table 3.19 Retention Index (RI) based on 1-nitropyrene
Sample RI for 1-nitronaphthalene RI for 2-methyl-1- RI for 2-nitronaphthalene nitronaphthalene
Standard Mix 0.528 0.545 0.551
LS&L-NO, 0.528 0.543 0.551
LS&M-NO 0.529 0.544 0.552
LS&H-NO 0.528 0.543 0.551
MS&L-NO,. 0.530 0.546 0.552
MS&M-NO, 0.528 0.545 0.551
MS&H-NO 0.528 0.544 0.551
HS&L-NO, 0.529 0.544 0.552
HS&M-NO 0.528 0.544 0.551
HS&H-NO 0.528 0.543 0.551
Key: L = low M= medium H=high S =speed
(RI taken from I injection of one of each duplicates)
106
lOO
100
Key 1 = 1-Nitronaphthalene 2 = 2-Methyl-1 -nitronaphthalene 3 = 2-Nitronaphthalene
10 20 25 30
2 1 22
40
23
Retention Time (minutes)
(a) Overall Sample
50 60
(b) Sub-section of Sample
(c) Sub-section of Sample + Nitro-PAC standards
24 25
Figure 3.20 The confirmation of the nitronaphthalenes in the mononitro-PAC HPLC fraction collected at high load and 1500 rpm (a) by comparing the nitronaphthalene temperature window (b) with that of the same sample eo-injected with the mononitro-PAC standard mix (c) (Chromatographic conditions given in Section 3.4.3.1.4.1).
107
Quantification was based on external calibration using the mononitro-PAC mix. Figure 3.2I
shows the calibration curve for I-nitronaphthalene and I-nitropyrene. Each mononitro-P AC
HPLC fraction was then injected twice (0.5 Ill). The calibration was checked following
completion of the injection of samples from I500 rpm and 2500 rpm. The mononitro-P AC
emissions were calculated in terms ofthe weight ofnitro-PAC emitted divided by the TES weight
collected and relative to the total fuel consumed during sampling.
3 4 3 I 4 2 Analysis ofDinitro-PAC HPLC Fractions by GC-ECD
This section describes the analysis of the dinitro-PAC HPLC fractions by GC-ECD. The
instrumentation and set-up parameters were as for those used for the analysis of the mononitro
p AC HPLC fractions (Section 3 .4.3 .l.4.I), with the following exceptions. The temperature
programme was initially 60"C held for I min, temperature ramp of I O"C/min to 180"C, then a
temperature ramp of 5"C/min to 300"C, held for I 0 minutes. The integration areas were defined
by the standard dinitro-PAC mix: I7-23 minutes for I ,5-dinitronaphthalene and I,8-
dinitronaphthalene, 28-30 minutes for 2,7-dinitrofluorene, and finally 32-36 minutes for the
dinitropyrenes (Figure 3.22).
The detection limits were: 94 pg for 1.8-dinitronaphthalene, 660 pg, for 2,7-dinitrofluorene, and
1.2 ng for I,8-dinitropyrene (SIN of31l). The analysis concentrated on the duplicate samples
taken at 3500 rpm. Samples were blown down to a constant weight under a gentle stream of
nitrogen. Samples were diluted to a concentration of2.5 mglml and injection volumes were 0.5
Jll. Identifications were attempted by eo-injection with the dinitro-PAC standard mix.
3 4 3 I 4 3 Analysis ofNitro-mcy-PAC HPLC Fractions by GC-ECD
This section describes the analysis of the nitro-oxy-PAC HPLC fractions by GC-ECD. The
instrumentation and set-up parameters were as for those used for the analysis of the mononitro
PAC HPLC fractions (Section 3.4.3.1.4.I), with the following exceptions. The temperature
programme was initially 60"C held for I min, temperature ramp of I O"C/min to 200"C, then a
temperature ramp of 5"C/min to 300"C, held for I 0 minutes.
108
o 1-nitronaphthalene y = 60.694x - 0.0451 R-sq = 0.997
140
120
100
(U
~ 80 <(
~
: 60 a..
40
20
0.0 0.5 1.0
ll. 1-nitropyrene y = 38.037x - 0.645 R-sq = 0.997
1.5 2.0
Weight Injected (ng)
Note: Duplicate injections at each weight
Figure 3.21 The calibration curves for 1-nitronaphthalene and 1-nitropyrene using gas chromatography with electron capture detection.
109
100 -
Key 1 = 1 ,5-Dinitronapbtbalene 2 = 1 ,8-Dinitronapbthalene 3 = 2,7-Dinitrofluorene 4 = 1 ,3-Dinitropyrene 5 = 1,6-Dinitropyrene 6 = 1 ,8-Dinitropyrene
....... ""-
I
10
2
4
3
u_l -T
20 30
Retention Time (minutes)
5
6
I~
40
Figure 3.22 The gas chromatogram of the dinitro-PAC standard mix. GC conditions: Supelchem PTE-5 column, 60°C for 1 min., 10°C/min to 180°C, 5°C/min to 300°C, held for 10 min.
110
The detection limits (SIN, 3/1) were: 312.5 pg for 2-nitrofluoren-9-one and 2.5 ng for 2,4,7-
trinitrofluoren-9-one. Samples were blown down to a constant weight under a gentle stream of
nitrogen. Samples were diluted to a concentration of2.5 mg/ml and injection volumes were 0.5
Ill. The integration areas were defined by the standard nitro-oxy-PAC mix: 20-23 minutes for
2-nitrofluoren-9-one, and 30-35 minutes for 2,7-dinitrofluoren-9-one and 2,4,7-trinitrofluoren-9-
one (Figure 3.23). It was found that the injection of the nitro-oxy-PAC HPLC fractions rapidly
caused the detector cell to become contaminated. For this reason, one of each of the duplicates
from the samples at 3500 rpm were injected. Identifications were attempted by eo-injection with
the nitro-oxy-PAC standard mix.
3 4 3 I 5 Gas Chromatography/Mass Spectrometry operated in the Negative Ion Chemical
Ionisation mode
Gas chromatograph/mass spectrometry operated in the negative chemical ionization mode
(GC/MS NICI) is very sensitive to electrophilic compounds such as nitro-PAC (Tong et al. 1984,
Ramdahl & Urdal 1982, Sellstrom et al. 1987, and Bayona et al. 1988).
The evaluation of GC/MS in NICI mode was performed on a Carlo Erba 5160 GC linked to a
Finnigan 4500 El - Cl mass spectrometer. Isobutane was used as the reagent gas. The source
temperature was 150°C, electron beam set at 45 eV, source pressure at 0.4 Torr, vacuum at 2.4-
2.8 x 10-s Torr. The Supelchem PTE-5 capillary column (25 m x 0.32 I.D., 0.25 llm film
thickness) with cooled on-column injector were used with the same temperature program as used
for the GC-ECD analysis. The output of the mass spectrometer was managed by a Data General
Nova/4 and INCOS computer.
Initially, the system was used in the full scan mode, however large source interferences were
evident, and the mode was switched to selective ion monitoring for specific molecular ions of
nitro-PAC. This also greatly increased the sensitivity, resulting in detection limits in excess of
those obtained with GC-ECD (Figure 3.24). The mononitro-PAC HPLC fraction from TES
collected at high load and 3500 rpm was analysed and the presence ofnitro-PAC was confirmed
by molecular ion and RI on 1-nitropyrene (Figure 3.25).
111
Key 1 = 2-Nitrofluoren-9-one 2 = 2,7-Dinitrofluoren-9-one 3 = 2,4,7-Trinitrofluoren-9-one
x = Septum bleed
100
3
X
2
10 20 30 40
Retention Time (minutes)
Figure 3.23 The gas chromatogram of the nitro-oxy-PAC standard mix. GC conditions: Supelchem PTE-5 column, 60°C for 1 min., 10°C/min to 200°C, 5°C/min to 300°C, held for lO min.
112
G.) (,J
= ~ "0 = = .c <
G.)
100
.~ 100 -~ ~
100
Key
1 = 40 pg of 1-Nitronaphthalene 2 = 40 pg of 2-Nitronaphtha1ene 3 = 40 pg of 2-Methy1-1-nitronaphtha1ene 4 = 52 pg of 3-Nitrofluoranthene 5 = 40 pg of 1-Nitropyrene 1
2
3
21.40 23.20 25.00
Retention Time (minutes)
4
5
39.10 40.00 40.50
Retention Time (minutes)
(a) 173 ion scan
(b) 187 ion scan
26.40
(c) 247 ion scan
4 1.40 42.30
Figure 3.24 The sensitivity of gas chromatography/mass spectrometry operated in the negative ion chemical ionisation and selective ion modes to a) 1- & 2-nitronaphthalene, b) 2-methyl-1-nitronaphthalene, and c) 3-nitrofluoranthene and 1-nitropyrene. GC conditions: Supelchem PTE-5 column, 60°C for 1 min., 5°C/min to 300°C, held for 10 min.
113
100
8.20
lOO
30.00 31.40
(a) Overall Sample, Selective Ion Mode (SIM): 173, 187, 191,205, 211,223, 225, 237, and 247 ions scanned
Retention Time (minutes)
- 1-Nitropyrene
33.20 35.00
(b) Sub-section of sample, 24 7 ion scanned
36.40 38.20
Retention Time (minutes)
Figure 3.25 The analysis of the mononitro-PAC HPLC fraction obtained at 3500 rpm and high load by gas chromatography/mass spectrometry operated in negative ion chemical ionisation mode, using selective ion monitoring (a) and the confirmation of 1-nitropyrene by scanning for the 247 molecular ion (b). GC conditions: Supelchem PTE-5 column, 60°C for 1 min. , 5°C/min to 300°C, held for 10 min.
114
3 4 3 2 Gas Chromatography/Mass Spectrometzy operated in Electron Impact mode for 01cy-P AC
Several researchers have utilized GC/MS in the El mode for identifying prominent oxy-PAC in
diesel extracts (Ramdahl 1983, Ktinig et al. 1983, and Schulze et al. 1984).
The HPLC fractions corresponding to the retention windows for dinitro-P AC, nitro-oxy-PAC,
and polar oxy-PAC were investigated for the presence of oxy-PAC by GC/MS in El mode. The
polar fractions from the silica gel chromatography from exhaust samples were also analyzed by
GC/MSEI.
A Rtx-5 (25 m x 0.32 ID, 0.25 J.lm film thickness) capillary column was used with cooled on
column injector and helium carrier gas. The temperature programme was 60°C for 1 min., 5°C/rnin
to 300°C, held for 10 rnins. The ionisation was 40 eV, with 40-544 amu scanned every 2 seconds.
The dinitro-PAC and nitro-oxy-PAC HPLC fractions had previously been diluted and injection
volumes were 1.0 J!l. The polar fractions from the silica gel clean-up were diluted with acetonitrile
to 5 mg/ml and 0.5 J!l injections were made. The samples were eo-injected with a P AH retention
index (0.3 J!l injected) containing naphthalene, phenanthrene, chrysene, and benzo(ghi)perylene.
The oxy-PAC standard mix was also eo-injected with the PAH retention index (Figure 3.26). The
retention index was used to aid identification of compounds.
3 4 3 3 Analysis ofn-Alkanes by Gas Chromatography with Flame Ionisation Detection
This section details the analysis of the aliphatic fractions collected at 1500 rpm and 3 500 rpm by
GC-FID. The ODS clean-up of the samples collected at 2500 rpm resulted in retaining the
aliphatic fractions on the cartridges. Following the work-up and isolation of the aliphatic fraction
from TES (Section 3.4.2), the aliphatic fractions were blown down to a constant weight. The
samples (1 0 mg/ml in DCM) were analyzed on the Carlo Erba Mega GC-FID fitted with a DB-5
capillary column (30 m x 0.32 mm, 0.25 J.lm film thickness, J&W Scientific), with hydrogen as the
carrier gas. Injection volumes were 0.5 J!l and cooled on-column injection was used. Integration
was by a Shimadzu integrator. The analysis and calibration procedures correspond to those used
for the aliphatic fractions from 1000 rpm, 2000 rpm, and 3000 rpm (Section 3.2.3}.
115
Key
Retention Index: A = Naphthalene B = Phenanthrene C = Chrsyene D = Benzo(ghi)perylene
100
A
3.38
j I 11.00
B
1
4
3
Oxy-PAC: 1 = Naphthaldehyde 2 = 9-Hydroxyfluorene 3 =Xanthone 4 =Anthrone 5 = Phenanthrene-9-carboxaldehyde 6 = 9-Phenanthrol 7 = 9-Fluorene-9-carboxylic acid 8 = Benzo(a)anthracene-7,12-dione
5
Total Ion Chromatogram (TIC)
6 7
c
2
8
D
'-.) '-J "- \Jl ~
'-.....J ...... '-.1\_ ..
18.22 25.44 33.06 40.48 47.50
Retention Time (minutes)
Figure 3.26 Total ion chromatogram of the oxy-PAC standard mix. GC conditions: Rlx-5 column, 60°C for 1 min., 5°C/min to 300°C, held for 10 min.
116
The n-alkane quantifications were based on heptadecane (R-sq = 0.9997). The stability of the
calibration was confirmed after the 1500 rpm samples. Results were calculated relative to the fuel
(Section 4.2.1). The aliphatic fractions from the fresh oil and sump oils were also analyzed.
3 4 3 4 Analysis of Major P AC by Gas Chromatography/Mass Spectrometzy operated in the
Electron Impact mode
Prior to analysis the engine samples collected at 1500 rpm, 2500 rpm, and 3 500 rpm were
concentrated, cleaned-up (Section 3.4.2) and the major PAH were isolated by HPLC fractionation
(Section 3.4.3). An aliquot of fuel (20.6 mg), taken during the nitro-PAC profiling sampling
sessions, plus d8-naphthalene (110 J.lg, 98+%, Aldrich Chemical Co.) and d10-phenanthrene was
(86 J.lg, 97%, Aldrich Chemical Co.) was subjected to silica gel clean-up (Section 3.4.2.2.2).
The analysis and identification of major P AH in the exhaust and fuel samples followed the same
instrument and set-up conditions as used for the major organics profiling session (Section 3.2.4),
with the following exceptions. Quantification was based on the response to deuterated fluorene
(0.228 mg/ml, Janssen Chimica), added as a quarter of the GC/MS dilution concentration (5
mg/ml) to all the samples (Table 3.20). Thus for a 0.5 J.ll injection 28.5 ng of deuterated fluorene
was injected. Results were calculated relative to the fuel (Sections 4.2.1 & 5.2.1.6). The aromatic
fractions from the fresh oil and sump oils were also analyzed.
Table 3.20 Response factors (R..) for PAC based on d,n-Fiuorene
Compound R,#l R,#2 R,#3 Aveml!e R,
Naphthalene 0.80 0.755 0.790 0.784
Dibenzothiophene 1.062 1.025 1.022 1.036
Phenanthrene 0.836 0.710 0.716 0.721
l'yrene 0.706 0.721 0.685 0.704
117
Chapter 4 - Primary Organic Emissions
This chapter examines the primary emissions from diesel combustion for a variety of engine
speeds and loads. The primary emissions are defined as being straight chain n-alkanes in the
aliphatic fraction and the major PAC, such as 2-4 ring PAH, in the aromatic fraction.
The establishment of the parameters controlling the primary organic emissions is vital in
relation to the improvement of diesel combustion with an aim of meeting proposed future
particulate limits and possible future PAH legislation. The origins of the organic emissions is
also fundamental to the elucidation of the formation processes resulting in secondary emissions
of highly mutagenic nitro-PAC.
The chapter is divided into four parts. Section 4.1 examines the relevant contributions of fuel
and oil survival to the Prima emissions. The results from samples collected with the initial and
upgraded TESSA configurations are presented and discussed in Sections 4.2 & 4.3
respectively. Finally, the combustion parameters responsible for the primary emissions are
summarized in Section 4.4.
4.1 The Contributions of Fuel and Oil Survival to the Emissions
Many diesel exhaust studies have shown that survival of lubricating oil and fuel occurs to
different degrees depending on the engine running conditions (Mayer et al. 1980, Cartellieri
& Tritthart 1984, and Cuthbertson er al. 1987). This section examines the relative
contributions of unbumt fuel and oil to the Prima emissions over a range of speeds and loads.
Figure 4.la shows a typical chromatogram of the aliphatic fraction of the standard diesel fuel.
The fuel is dominated by an homologous series of n-alkanes ranging from C9 to C28, with the
maximum n-alkane concentration associated with hexadecane (Table 4.1). The aliphatic
fraction of lubricating oil consists largely of an unresolved complex mixture (UCM) eluting
at high temperatures, with a series of high boiling n-alkanes superimposed on the UCM hump
(Figure 4.1b).
118
100 -
Cl2
c~o I I. J
I
100 100
100 - lOO
I lt l I
100
Cl6
Cl8 Cl4
C20
C22
J '~ " " 11. ., .. .1
IC24
I I I
I 200
200
Cl6
JJJ LA I ill . . l _I I I I
I 200
Temperature ( ° C)
I 300
300
I 300
(a) Fuel
(b) Lubricating
Oil
(c) TES
Figure 4.1 Comparison of the aliphatic fractions of diesel fuel (a) and lubricating oil (b), with the TESSA extracted sample (TES) at 1% of full load and 1000 rpm (c) (Chromatographic conditions given in Section 3.2.3).
119
A comparison of the fuel and oil with the aliphatic fraction of a TES collected at low speed
and low load is shown in Figure 4.lc. The compounds associated with the fuel temperature
region are clearly visible, whereas the lubricating oil hump is indiscernible.
Table 4 l Concentration of Selected n-alkanes in the A2 Fuel
Compound Relative molecular Melting point ("C), Concentration mass Boiline ooint ("C)' {J)J)IIl)
Decane lClO) 142 "29.7, 174.1 1578
Dodecane (Cl2) 170 "9.6, 216.3 3924
Tetradecane lCI4) 198 5.9, 253.7 8855
Hexadecane (C16) 226 18.2, 287.0 14529
Octadecane (C 18) 254 28.2, 316.1 10369
Eicosane CC20) 282 36.8, 343.0 7542
Docosane lC22) 310 44.4, 386.6 4702
Tetracosane (C24) 338 54.0, 391.3 2697
Hexacosane (C26) 366 56.4, 412.2 1030
Octacosane lC28) 394 64.5, 431.6 244
1 Weast & Astle (1985)
After enlarging the lubricating oil temperature window there is evidence of a small oil hump
eluting between 275"C and 300°C (Figure 4.2). This hump does not cover the entire range of
the lubricating oil UCM (165°C to 300'C), suggesting that only the very high molecular weight
constituents of oil leakages survive unchanged. The great majority of any lubricating oil
leakages are combusted. There does appear to be some peak matching of the exhaust to the
lubricating oil throughout the lubricating oil range. This shows that specific components of the
oil may have sufficient resistance to combustion, allowing these compounds to survive
unchanged. The absence of a significant oil UCM in the aliphatic fraction of the exhaust at low
load and low speed, indicates that even at the engine conditions known to favour oil survival
(Williams et al. 1987), there are only minor oil survival contributions to the Prima exhaust.
The oil contributions remained low for increased loads and speeds (Figure 4.3). For all the
engine samples collected, no significant oil UCM could be detected, even after reducing the
chart recorder attenuation within the oil temperature zone.
120
j
Key
+ = matching of compounds in both samples
200
C2P
' J..
I 200
300 Temperature ( OC)
+
(a) Lubricating
Oil
(b) TES
~~ ++
lA lt u . L J..» 1 ~ -""'-
I I 300
Temperature ( OC)
Figure 4.2 Close examination of the oil hump temperature window (a) to the same temperature range of the TESSA extracted sample (TES) at 1 % of full load and 1000 rpm (b) (Chromatographic conditions given in Section 3.2.3).
121
I
100 Cl6
100 2 0
~ ~
= C16 c:> Cl. :3 ~
~
100 200
C l6 ~ ~
= c:> Cl. :3 ~
~
200
Temperature ( °C)
(a) Mid Load
& Low Speed
300
(b) High Load
& Low Speed
300
(c) Low Load
& High Speed
300
Figure 4.3 Examination of the aliphatic fractions of TESSA extracted samples (TES) for evidence of oil contributions at three power settings: 50% of full load & 1000 rpm (a) , full load and 1000 rpm (b) , and 1% of full load and 3000 rpm (c) (Chromatographic conditions given in Section 3.2.3).
122
Figure 4.4 shows the gas chromatographic traces of the aromatic fractions of the A2 fuel,
lubricating oil, and exhaust at low speed and load. Figure 4.4a and Table 4.2 indicate that the
fuel is dominated by the smaller 2-3 ring PAH and the corresponding methyl derivatives.
Further evidence that fuel survival is the dominant source is shown by the close chemical
resemblance of the fuel and emissions character. The high boiling aromatic constituents of the
oil are not recovered at significant levels in TES.
Table 4 2 Concentration of Selected PAC in the A2 fuel
Compound PAC Relative Melting point ("C), Concentration class molecular Boiling point ("C)' (ppm)
mass
Naphthalene PAH 128 81, 218 2593
Methylnaphthalenes PAH 142 22-35, 245-251 8632
D imethy lnaphtbalenes PAH 156 1.6-108, 262-276 18432
Fluorene PAH 166 115-116, 294 831
Methylfluorenes PAH 180 46-87, 302-318 2116
Dibenzothiophene PASH 184 97, 3322 296
Methyldibenzothiophenes PASH 198 66-85,- 634
Phenanthrene PAH 178 101, 338 1631
Methylphenanthrenes PAH 192 52-123, 352-359 3980
Pyre ne PAH 202 151, 393 151
Chrysene PAH 228 254, 431 31
1 Karcher et al. (1985) 2 Weast & Astle (1985)
Although the contributions of oil survival to TES were low, the accumulation of unbumt fuel
components in the sump was investigated. Figure 4.5 shows there was no or very little build
up of the lighter fuel aliphatics in the sump oil after 50 and 105 hours of usage. Survival of
used oil is not responsible for aliphatic emissions eluted within the fuel hump retention zone.
The accumulation of aromatics in the sump oil was also examined (Figure 4.6). There was a
moderate build-up of PAH as can be seen by the appearance of phenanthrene and the methyl
phenanthrenes in used sump oil. After adjusting for laboratory losses it can be shown that the
level of PAH in the oil increased with oil usage (Figure 4. 7).
123
lOO
lOO
100
Key I = Naphthalene
Total Ion Chromatogram (TIC): 2
2 = Methyl-naphthalenes 3 = Dimethyl-naphthalenes 4 = Trimethyl-naphthalenes 5 = Phenanthrene
TIC
~ 6 = Methyl-phcnanthrenes
a = d 8
-naphthalene
b = d1 0 -fluorene
c = d 10
-phenanthrene
(a) Fuel
7.23 14.49 22.15 29.41 37.07 44.33 51.99
a b c (b) Lubricating
Oil
7.23 14.49 22.15 29.41 37.07 44.33 51.99
(c) TES
7.23 14.49 22.15 29.41 37.07 44.33 51.99 Retention Time (minutes)
Figure 4.4 Comparison of the aromatic fractions of fuel (a) and lubricating oil (b), with the TESSA extracted sample (TES) at 1% of full load and 1500 rpm (c) (Chromatographic conditions given in Section 3.2.4).
124
100
100 100
lOO 100
100
200
200
L
200 Temperature ( ° C)
300
300
300
(a) Fresh Oil
(b) Sump after
50 hours
(c) Sump after 105 hours
Figure 4.5 Examination of the aliphatic fractions of sump oils for accumulation of unburnt fuel. Fresh oil (a), sump after 50 hours use (b) , and sump after 105 hours use (c)
(Chromatographic conditions given in Section 3.2.3).
125
100
100
~ "J = = c. fj ~
9 ~
100
Key Ph = Phenanthrene Me-Ph's = Methyl-pbenanthrenes
dNa dFl dPh
100 200
Ph
I TPh'•
100 200
Internal standards:
dNa = d g -naphthalene dPh = dto·phenanthrene dFl = dto-fluorene
300
300
(a) Fresh Oil
(b) Sump after
50 hours
(c) Sump after 105 hours
Temperature ( ° C)
Figure 4.6 Examination of the aromatic fractions of sump oils for accumulation of unburnt fuel. Fresh oil (a), sump after 50 hours use (b), and sump after 105 hours use (c) (Chromatographic conditions given in Section 3.2.4).
126
25 D Phenanthrene 1t.. 2- Methylphenanthrene V 3- Methylphenanthrene -C) 20 0
~ ........ C)
E 0 - 15 V c:
0 V 4J
cu .... 10 a 4J
c: Q) 0 c: 0 5 ()
0
0 20 40 60 80 100 120
Oil usage (hours]
Figure 4. 7 Accumulation of phenanthrene and methyl derivatives in sump oil over time.
127
Previous studies have shown an accumulation of both lower n-alkane and PAC in the sump oil
over a period of time (Abbass et al. 1987). Abbass and eo-researchers found 30 mg of
phenanthrene/kg of oil had accumulated after 106 hours for a 4 1 Perkins DI engine. The
authors found the accumulation levels reached an maximum after 120 hours of oil use.
Compared with the study by Abbass et al. (1987), the accumulation of phenanthrene in the
sump oil for the Prima engine was lower, corresponding to ea. 20 mg/kg after 105 hours of
oil use.
The lack of prominent peaks present in the aromatic fraction of the oil sump, collected after
50 hours of use, is of interest. Compared with the other two oil samples, the sump sample
collected after 50 hours of use was subjected to HPLC fractionation to examine the nitro- and
oxy-PAC content (Section 3.4.2.3). Absence of prominent peaks indicates that there are more
polar constituents in the oil which elute after the PAH fraction. The composition of the more
polars fractions of the oil are discussed in Chapter 5.
An examination of the relative contributions of fuel and oil at varying engine conditions,
reveals that fuel survival is by far the dominant survival source. The results presented later in
this chapter were derived from two sampling sessions. For both sessions the oil usage over the
sampling periods was low, and as a result the accumulation of organics is correspondingly low.
This allows emissions from the earlier samples to be directly compared to the later samples.
Finally, the minor contribution of lubricating oil toTES means that the limited accumulation
of n-alkane and PAC which does occur will not add significantly to the Prima exhaust.
4.2 Profiling of the Primacy Organics using the initial TESSA configuration
Engine emissions are strongly influenced by driving conditions and cycles, for example the
combustion environment at high speed motorway driving will be different from that evolved
at urban congested driving. The contribution of survival and pyrosynthesis/pyrolysis to
emissions may also be affected by differing engine conditions. Engine sampling under different
conditions followed by comparing the chemical proftles or 'fingerprints' produced at the
different conditions, allows an estimation of both the effect of engine conditions on the
emission levels and also the dominant emission pathways.
128
The results in this section were obtained by sampling at 1000 rpm, 2000 rpm, and 3000 rpm
using the initial TESSA set-up (Section 3.2.1). The samples were then extracted, concentrated,
and fractionated by silica gel open column gravity fed chromatography (Section 3.2.2). The
n-alkanes and PAH were quantified by GC-FID (Section 3.2.3) and GC/MS El (Section 3.2.4)
respectively.
4.2.1 The Concept of Percentage Recoveries
The major pathways for the source of the primary organic emissions, ie. compounds present
in the fuel and/or oil, in the Prima exhaust are:
a) survival of the compound from the fuel/lube oil unchanged
b) combustion reactions in which fragments of fuel/lube oil pyrosynthesize, resulting
in partially combusted products.
c) transformation reactions in the exhaust pipe.
In (a) the carbon skeleton of the compound remains unchanged, while in (b) the carbon
skeleton of the compound is altered/newly formed in the combustion chamber. The results
presented in Section 4.1 identified that lubricating oil contributions to TES are minimal in
comparison to those derived from fuel survival. The close proximity of TESSA to the engine
will reduce the opportunities for post-combustion exhaust reactions (c). Therefore, the
following results and discussions are focused on the survival and combustion products of diesel
fuel.
For different compounds in the emissions, the relative contributions by mechanism (a) may
vary. This will reflect the different combustion efficiencies that occur for differing organic
structures under different combustion conditions. The similarity in ratios of compounds in the
fuel and the emissions may thus be a measure of the conditions which favour similar
combustion efficiencies, so that they survive to similar degrees. By contrast, very different
ratios in emissions compared with the fuel, may indicate conditions which favour differences
in degree of destruction of structures. Interpretation here is made more complicated because
different ratios may also indicate a significant contribution from pyrosynthetic reactions (ie.
mechanism (b)).
129
The degree of combustion for a compound is related to a number of factors in the combustion
chamber, eg. temperature, oxidant concentration, air motion, and kinetics of reaction
(Campbell et al. 1981, Scheepers & Bos 1992b). For a molecule within the combustion
chamber to react, it may be necessary for it to experience a sufficiently high temperature to
break bonds, sufficient oxidant and a sufficient length of time to allow the reaction to be
completed. The extent of reaction that occurs in the combustion chamber will depend
especially on temperatures and the kinetics of the reaction.
In an attempt to evaluate the relative contributions of the emission pathways, namely survival
and pyrosynthesis/pyrolysis, at varying engine conditions, the recovery of primary emissions
as a percentage of the corresponding fuel compound supplied to the combustion chamber
were calculated. An example of the percentage recovery of pyrene at 1% of full load and
1000 rpm is given below:
py PY - ex .rtOO
PlC pyfu
= 0. 0018 .A1.00=1.16% 0.155
where PY PR = percentage recovery of pyrene PYex = pyrene in exhaust sample collected (mg) PYru = pyrene in fuel consumed during sampling period (mg)
The comparison of the percentage recovery of different compounds at a particular engine
condition enables an insight into the source of the emissions. Figure 4.8 uses a theoretical
model to demonstrate the interpretation of the percentage recoveries. If all compounds (A,
B, and C) are recovered by the same amount, then the ratio of these compounds in the
exhaust matches that in the fuel, ie. fuel survival unchanged. Differing percentage recoveries
may indicate that the structure of compound C results in a greater combustion compared with
compound A. On the other hand, a proportion of compound A may be added to by
pyrosynthetic reactions, such as compound C being transformed into compound A.
130
1.2
1
0.8
0.6
0.4
"0 0.2
c :::1
0 8. E 0 u 4-o 0
c <1) ;;>
1.2 0 (,J <1)
~ <1) 1 0.0 ~ ..... c 0.8 <1) (,J I-.
~ 0.6
0.4
0.2
0
A
A
B Compound
•·· .
c
.......................................
B c Compound
~ FUEL SURVIVAL ..,........... UNCHANGED
PREFERENTIAL SURVIVAL
COMBUSTION GENERATION
Figure 4. 8 Interpretation of percentage recovenes for identification of the source of emiSSions.
131
The percentage recovery for selected PAC and n-alkanes for a range of speeds and loads are
shown in Tables 4.3 and 4.4 respectively. To aid the comparison of the results to other
combustion studies, the data is also presented in terms of the amount of compound emitted
relative to the total amount of fuel consumed (units of mg/kg).
Table 4.3 The Exhaust PAC emissions relative to the corresponding fuel PAC consumed during sampling
Engine Percentage Fluorene D ibenzothiophene Phenanthrene Pyre ne Speed of Full (rpm) Load
1000 1 0.91 (6.83) 0.88 (3.28) 0.86 (14.28) 1.16 (1.79)
25 0.26 (1.95) 0.20 (0.76) 0.20 (3.31) 0.14 (0 .22)
50 0.19 (1.45) 0.17 (0.63) 0.16 (2.72) 0.13 (0.21)
75 0.20 (1.50) 0.18 (0.65) 0.17 (2.75) 0.15 (0.23)
lOO 0.21 (1.57) 0.20 (0.75) 0.21 (3.49) 0.19 _{0.31)
2000 1 0.36 (2.70) 0.31 (1.14) 0.29 (4.80) 0.30 (0.47)
25 0 .32 (2.44) 0.23 (0.86) 0.18 (3.60) 0 .14 (0.21)
50 0.21 (1 .55) 0.16 (0.59) 0.16 (2.67) O.ll (0.16)
75 0.22 (1.65) 0.18 (0.68) 0.16 (2.70) 0. 10 (0.15)
100 0.16 (1.19) 0.15 (0.56) 0.15 (2.56) 0.08 (0.12)
3000 1 0.63 (4.72) 0.49 (I .83) 0.45 (7.52) 0.29 (0.44)
25 0.46 (3 .44) 0.35 (1 .29) 0.26 {4.38) 0.19 (0.30)
50 0.31 (2.31) 0.28 (1.06) 0.23 (3.82) 0.13 (0.20)
75 0 .17 (1.25) 0.19 _(0. 70) 0.16 (2.71) 0.09 (0.14)
lOO 0.13 (0.99) 0.17. (0.64) 0. 13 (2.19) 0.08 (0. 12)
Figures in brackets expressed as compound (mg) recovered per kg of fuel consumed (mg/kg).
Emissions expressed relative to the total fuel consumed take no account of the varied
concentrations of compounds in the fuel itself. This can translate to compounds having the
same percentage recovery at specific conditions but different values when expressed relative
to the total fuel burned. For example fluorene and phenanthrene are both recovered by 0.21 %
at 1000 rpm and full load. However, the greater phenanthrene compared with fluorene in the
132
fuel, results in higher emissions relative to the total fuel consumed for phenanthrene.
Expressing the emissions as a percentage recovery allows the combustion of compounds at
differing fuel concentrations to be directly compared.
Table 4.4 The Exhaust n-alkane emissions relative to the corresponding fuel n-alkane consumed during sampling
Speed Percentage Cl4 C15 C16 C18 C20 (rpm) ofFuU
Load
1000 1 0.55 (46.0) 0.51 (61.0) 0.39 (57.5) 0.43 (43.3) 0.50 (33.6)
25 0.17 (13.9) 0.14 (16.8) 0.1 (14.7) 0.09 (8.7) 0.09 (6.6)
50 0. 11 (9.5) 0.10 ( 12.6) 0.08 (11.6) 0.07 (7.3) 0.07 (5.0)
75 0.12 (10.2) 0.11 (13.4) 0.08 (11.9) 0.08 (7.7) 0 .07 (4.9)
100 0.10 (8.5) 0.10 (11.4) 0.07 (10.6) 0.08 (7.5) 0.07 (5.2)
2000 1 0. 15 (13.0) 0.18 (21.8) 0.12 (17.9) 0.09 (9.2) 0.08 (5.4)
25 0. 10 (8.3) 0.05 (5.7) 0.07 (9.6) 0.04 (4.3) 0.04 (2.5)
50 0. 11 (9.3) 0.11 (12.9) 0.07 (10.5) 0.05 (4.9) 0.04 (2.9)
75 0.06 (5.2) 0.06 (7.2) 0.04 (6.1) 0.03 (2.8) 0.02 (1.5)
100 0.09 (7.4) 0.10 (12.2) 0.07 (9.8) 0.05 (4.7) 0.03 (2.3)
3000 I 0.31 (25.7) 0 .29 (35. 1) 0.18 (26.8) 0.14 (13.6) 0 .13 (9.3)
25 0.19 (15.8) 0.18 (21.2) 0.11(16.2) 0.07 (7.3) 0.06 (4.4)
50 0.19 (15.5) 0 .19 (22.5) 0.13 (19.1) 0 .08 (8 .2) 0.06 (_4.1}
75 0.12 (9.7) 0.13 (15 .6) 0.10 (14.6) 0 .08 (7.5) 0.05 (3.6)
100 0.03 (2.3) 0.04 (4.2) 0.03 (4.8) 0.03 (3.2) 0.02 (1.7)
Figures in brackets expressed as compound (mg) recovered per kg of fuel consumed (mg/kg).
133
4 2.2 The Effect of Engine Load on the Percentage Recovery of Primary Organics
This section examines the effect of engine load on the combustion efficiency of specific
components of diesel fuel, namely those of selected n-alkanes and 2-4 ring PAC. Combustion
efficiency is thermodynamically defined as the actual energy produced from the fuel:air charge
relative to the theoretical maximum. Calculation of combustion efficiency defined then requires
complicated temperature measurements, which are particularly difficult with the heterogenous
nature of diesel combustion. Combustion efficiency can also be defined as the mass of fuel
emitted from the cylinder relative to the mass of fuel injected, ie. in the form of percentage
recovery. Whenever the term combustion efficiency is used in this study for a specific
compound it is expressed as the level of its percentage recovery. For example, a 0.1%
recovery is equivalent to 99.9% combustion efficiency.
Figure 4.9a illustrates the recovery of selected PAC in the emissions from an engine run at
1000 rpm and varying loads. For a specific load the closeness of the points for the different
PAC, indicates ratios of PAC content similar in the emissions to those in the fuel. This is
strong evidence for survival of the PAC unchanged (mechanism (a)). It is also apparent from
Figure 4.9a that the individual recoveries of PAC are lower at higher loads compared to the
lowest load. At the lowest load the total yield of PAC in the emissions is at its maximum and
the ratios are observed to be much more variable than at all other loads for this speed (pyrene
especially appears to be atypical at low load). The n-alkane molecules exhibit a similar
relationship with engine load as for the PAC emissions at low engine speed (Figure 4.9b). The
greatest recoveries are, as for the aromatics, associated with low loads. The closeness of the
percentage recovery of different n-alkanes at each load position again suggests that the organic
emissions are derived from fuel surviving combustion unchanged.
Further evidence for the PAC and n-alkane emissions being derived from fuel survival at low
speed is indicated by the decline of total unburnt hydrocarbon (UHC) and TES emissions with
engine load (Figure 4. lOa & b). The UHC and TES emissions relative to the total fuel
consumed both decrease with increased load. The aromatic and n-alkane emissions follow the
main emission pathway of the UHC and TES emissions, namely survival.
134
a)
1.2 o Fluorene
* Dibenzothiophene
1.0 A Phenanthrene
o Pyrene >-.... .,
0.8 > 0 0 ., a: ., 0.6 C) 10 ... c: .,
0.4 0 '-., Q.
0.2 ~ =--8 - ~
0.0 0 20 40 60 80 100
Percentage of Full Load
b)
1.20
1.00 o C14
>- o C15 '-., > 0.80
A C16 0 0 ., a: V C18 ., 0.60
0 C20 a «r ... c: ., 0 '-G)
a..
0.20
i 8 0.00
0 20 40 60 80 100
Percentage of Full Load
Figure 4. 9 Effect of engine load on the percentage recovery of selected PAC (a) and nalkanes (b) at LOOO rpm.
135
al
~ 150 • E 0 :::1 125
~ ... c: 0 u ... 100 :::1 -- 75 0~ 0
Ot -u 50 o......_ :I: 0----:::l 0 - 25 0
Ot
E 0 0 20 40 60 80 100
Percentage of Full Load
bl 30
~ • 0 E :::1 25 ... c: 0 u 20 ... :::1 -- 15 0
Ot - 10 Ot 0
.§. ~ ----0 0 5 Cl) 0 w 1-
0 0 20 40 60 80 100
Percentage of Full Load
cl 6
5 / a 0 z
4 • ....: 0
/0 E 3 Cl) 0 .s=
/ u 2 ... 0
ID
------0 0
0 0 20 40 60 80 100
Percentage of Full Load
Figure 4.10 Effect of engine load on the emissions of unburm hydrocarbons (UHC) (a) , TESSA extracted sample (TES) (b), and bosch smoke (c) at 1000 rpm
t36
Figure 4.10c shows the trend of smoke, measured by a Bosch smoke meter, relative to the
engine load. The Bosch smoke measurements can be equated to the dry soot composition of
particulates, since the technique relies on the absorption of light by the carbon nuclei, rather
than a weight determination (Alkidas 1984). A weight determination would include both 'dry'
and 'wet' organics. The trend for the smoke is the opposite to the organic emissions, such that
the soot output increases with increased load. This indicates that the exhaust composition at
low loads is enriched in organics whereas the dry soot contribution to total particulates
dominates at high loads. This agrees with the findings of Iida et al. 1986, who studied the
particulate and SOF emissions from a single cylinder 0.9 1 DI diesel engine. These authors
found that at full load, 95% of the total mass of particulates was dry particulate, whereas at
zero load 100% of the particulate was SOF based. Draper (1986) also found that under full
load the particulate collected was less than 3% SOF, whereas under reduced load the exhaust
particulate was 29% SOF. A similar fmding was found by Gomes & Yates (1992) for a single
cylinder 0.61 Hydra engine, for which the SOF contribution at low loads was between 80%
to 90%, whilst much reduced at high loads.
The average fuel survival for the selected PAC decreased from 0.95% at low load to 0.2% at
full load. The average fuel survival for selected n-alkanes ranged from 0.48% at low load to
0.084% at high load. This shows that fuel survival exhibited a negative association with engine
load. This finding is in agreement with previous studies (Table 4.5), which also suggested fuel
survival was highest at low engine loads. Other studies, such as that by Assaumi et al. 1992,
have found that combustion conditions deteriorated badly at low loads and led to greater SOF
emissions. Mills et al. (1984) also found the total PAH emissions were highest at low loads
for a light-duty DI diesel, suggesting inefficient fuel combustion under these conditions.
Barbella et al. (1989) also found that PAH emissions relative to the fuel consumed were
highest at low loads. The authors showed that different PAH exhibited varying trends across
the load range, such that fluorene and phenanthrene initially decreased at mid-load before
rising again at higher loads. By contrast, fluoranthene and benzo(e)pyrene progressively
declined with increased load. The size of the engine and the configuration of the combustion
chamber may be significant factors in the emission trends. For example, the study of a
medium-duty Perkins Dl 4-236 4 I diesel engine by Williams et al. (1989) found opposite
137
trends to the earlier study of a smaller 0.5 I per cylinder DI engine. For example the recovery
of phenanthrene at low, mid, and high load was 0.05%, 0.03%, and 0.58% respectively for
the 4 I diesel (Williams et al. 1989). The PAH emissions from IDI diesel engines are also
highest at high load (Zierock et al. 1983 and Abbass et al. 1991a). It is important to note, that
even within the same fuel injector type and engine capacity range, the combustion efficiencies
are greatly influenced by the actual design used (Kamimoto & Kobayashi 1991). Overall, the
limited studies of the primary organic emissions, relative to the fuel consumed, from light-duty
DI diesels suggest that the lowest combustion efficiencies are generally associated with low
loads.
Table 4.5 Previous Studies on Fuel Survival for Diesel Engines
Engine Type Speed Load Emission Levels Reference (rpm)
0121 1500 phenanthrene emitted/total Andrews et al. fuel burned: 1983
Low 0.15 mgfkg Mid 0.06 mg/kg Hi~h 0.04 m~/k~
DI 1 cylinder 1500 Recovery of Pyrene: Williams et al. 0.5 1 Low 0.8% 1986
Mid 0.2 % ffigh 0.2%
Heavy-duty DI 2000 air:fuel ratios: Pyrene em.itted!total fuel Barbella et al. turbocharged burned: 1988 run on h.igh 23 0.6 mg/kg aromatic fuel 44 2.5 mg/kg
DI 1 cylinder 3600 % of full load: Total PAH emitted Ziejewski et al. 0.51 p.g/; of fuel: 1991
1 3.1 25 2.22 50 0.38 75 0.22 100 0.81
Engine load is strongly correlated with combustion chamber and exhaust temperatures (Table
4.6) . Since engine oil and water remove heat from the engine, the values of these two solutions
can be used to follow the temperature of the Prima engine. Referring to sampling details
tabulated in Section 3.2.1 indicates temperatures of both oil and water, and hence combustion
temperatures, rise with engine loads.
138
Table 4.6 The effect of engine load on the combustion and exhaust temperatures
Engine Speed Load Combustion Exhaust References Type chamber temperature ("C)
temperature ("C)
1 cylinder - Zero 190 Jensen & Half - 470 Hites 1983 Full 590
1.212 2350 rpm Zero 178 Mills & cylinder Quarter - 232 Howarth DJ Mid 302 1984
Three quarters 408 Full 533
0.871 2400 rpm Low 500 lida et 1 cylinder Mid - 600 a/.1986 DI High 900
0.51 DJ 1500 rpm Low 146 Andrews et Mid - 208 al. 1987 High 371
0 .7 11 1500 rpm Low 80 Pipbo et al. cylinder Mid - 300 1991 DJ High 430
0.5 11 - Zero 250 to 300• Chang et al. cylinder Full 450 to 500 - 1993
a Temperatures taken 38° after TDC
The peak gas temperatures in the combustion chamber are much higher than the exhaust, with
typical combustion temperatures near TDC between 1000 to 2000°C (Chang et al. 1993). The
higher combustion temperatures associated with full loads increase the likelihood that an
organic compound will be completely combusted. It is important to note that the heterogenous
nature of diesel combustion will result in a range of temperatures for any given engine speed
and load. For example, Li 1982 found the temperatures (between 232°C and 252°C) in the
space between the top ring liner and piston head, termed the crevice volume, lower than the
temperatures encountered in the piston bowl (ea 280°C). Saito et al. 1986 showed that the
temperature of the top-edge of the combustion bowl wall was 237°C whilst only 215°C at the
bottom of the chamber. Signer & Steinke (1987) also found lower temperatures in the crevice
volume associated with the oil liner zone compared with the bowl edge, and that the higher the
top ring the less space available for the flame to penetrate, and hence lower temperatures.
139
A temperature gradient or profile, as used by Schmeltz & Hoffmann (1976) to describe the
different temperatures in cigarette combustion, can also be envisaged to describe the range of
temperatures at a particular engine condition. The overall temperature for the temperature
gradient will be lowest at low loads, and the average temperature within the gradient will
increase with engine load.
One of the largest contributors to UHC at light loads arises from excessive air:fuel ratios
taking the fuel from the spray edge beyond the lean flammability limit (Greeves 1979, Yu
1980, Campbell et al. 1981, Wheeler 1984, Gomes & Yates 1992, and Stone 1992). Low
engine load also gives rise to hydrocarbon emissions via 'quenching' of the flame front in the
clearance between the piston top and the cylinder head near TDC (Laity et al. 1973, Matsui
& Sugihara 1986, Kobayashi et al. 1992, and Horrocks 1993). The crevice volume also gives
rise to high UHC emissions, since the trapped gases at around TDC cannot be reached by the
flame during compression, and consequently the unbumt fuel is emitted during the expansion
stroke (Ferguson 1986 and Horrocks 1993).
Low combustion temperatures associated with low loads allows fuel impinged on the chamber
walls to survive combustion later in the cycle and become emitted (Yu et al. 1980, Haupais
1982, Lilly 1984, Andoh & Shiraish 1986, Tsunemoto et al. 1986, Gomes & Yates 1992,
Tanabe et al. 1992, and Cossali et al. 1993). A study of a Ricardo Hydra 0.6 I per cylinder
engine found fuel impingement on the bowl walls occurred under all loads (Rao et al. 1993),
as did the 4 I diesel studied using high-speed photography by Karimi (1989). It follows that
the small Prima engine may also encounter fuel impingement over a range of loads. The fuel
impingement would be greatest at high loads, since more fuel is injected later in the cycle.
However, the higher chamber temperatures and greater turbulence mixing at high loads may
account for a greater combustion of the fuel deposited on the walls. In some cases, fuel
impingement can increase turbulence and further promote the combustion of the fuel
(Kamimoto & Yagita 1989, Matsuoka 1990, and Kamimoto & Kobayashi 1991), especially
for engines specifically designed to utilize impingement (Lilly 1984 and Akaska & Tamanouchi
1992).
140
A major source of UHC is derived from the fuel contained within the residual sac volume
(RSV) of injectors and the fuel spray tail injected late in the cycle surviving combustion
(Henein 1973, Greeves 1979, Campbell et al. 1981, Kowalewicz 1984, Lilly 1984, Wheeler
1984, Andoh & Shiraish 1986, Beck et al. 1988, Gomes & Yates 1992, and Horrocks 1993).
The survival of the RSV and the spray tail will be increased by lower cylinder temperatures.
Similarly, uncontrolled fuel releases such as those associated with injector dribble and
secondary stage injection enter late in the combustion cycle and survive under low
temperatures (Henein 1973 and Stone 1992).
Iida & Sato 1988 studied the UHC and SOP emissions from 1 cylinder of a Dl diesel engine,
and found that the emissions (units of g emitted/kg of fuel) correlated negatively with
temperature at low loads. As the temperature decreased a linear increase in the SOF and UHC
was found.
Overall, the mixing beyond the lean flammability limit and low combustion temperatures
associated with low loads, favour condensation and adsorption of unbumt fuel from the
outlined sources onto soot particulates, resulting in large SOF contributions to total
particulates.
Figures 4.9a & b may be used to show the extremes of conditions related to temperature. At
high load the higher temperatures in the combustion chamber result in greater combustion
efficiencies for the selected compounds. The molecules of organics which are collected in the
emissions represent those which have escaped totally the combustion temperatures as indicated
by the closeness of their recoveries (similar ratios to fuel ratios). By contrast, at low load the
inefficient combustion results in lower temperatures and the local temperature gradients may
become more pronounced. The greater heterogenous temperature zones occurring at lower
loads may lead to a greater variation in the combustion efficiencies for different organic
compounds. At high load the temperature gradient may be narrower, as a consequence of
greater combustion turbulence, mixing the chamber temperatures to a greater extent than at
low loads (Shiozaki et al. 1980).
141
4.2 3 Effect of Engine Speed on the Percentage Recovery of Primary Organics
Figure 4.11a showing the full range of load at 3000 rpm indicates a wider separation of PAC
emissions/fuel ratios compared with lower speed runs. The average recovery at low load was
0.47% and this value declined progressively with increased load to 0.13% at full load.
Increased speed leads to increasing average efficiency of combustion of the PAC. The wider
range of ratios for a specific load, suggest differing combustion efficiencies between the PAC
at all loads, or alternatively combustion generation reactions have become more prominent.
The greatest variation between the PAC recoveries is again associated with the lowest load.
The percentage recoveries of the n-alkanes at high speed also decrease with increased engine
load (Figure 4.11b). The n-alkane percentage recovery are again lower than those of the
aromatics, with the average recovery decreased from 0.21% to 0.03% at full load. The wide
variations between different n-alkane percentage recoveries at low load is similar to the
aromatic emissions at low load. For both classes of organics the percentage recoveries
converge at high load, providing further evidence for fuel survival unchanged at high loads and
speeds. The closeness of the trends for the n-alkanes and PAC at all conditions suggests the
aromatic and n-alkane emissions are derived from the same sources.
The selected P AC and n-alkane emissions follow the TES and UHC trends with load at high
speed (Fig 4.12). The soot emission follows the same trend as for the lower speed, ie.
increased emissions with increased loads. All three types of emissions are at lower levels than
at the lower speed, showing the improved combustion at higher speeds.
Engine speed effects the air motion (swirl, squish, and turbulence) characteristics, injection
timing, and combustion temperatures of the engine (Jensen & Rites 1983, and Rao et al.
1993). The effect of air swirl, the dominant air motion parameter, is to increase the rate of air
entrainment into the fuel jet (Campbell et al. 1981, Cossali et al. 1993, and Singal et al.
1993). The tangential motion of the swirl tends to deflect the spray penetration of the chamber,
whilst improving the atomization and vaporization. Increased swirl may also further the
turbulence and thereby promote fuel-air mixing in the later stages of combustion (Shioji et al.
1989).
142
>.... CD > 0 u CD a:
CD 0) ftl -c CD u .... CD
D..
a) 1.2
1.0
0.8
o Fluorene
* Dibenzothiophene
A Phenanthrene
o Pyrene
0 .0 --'-r---.------.--.-------,.----,
>.... CD
1.20
1.00
> 0.80 0 u CD a:
CD 0 .60 0) ftl -c CD ~ 0.40 CD
D..
0 .20
0
b)
0
20 40 60 80
Percentage of Full Load
20 40 60
o C14
D C15
A C16
V C18
~ C20
80
Percentage of Full Load
100
100
Figure 4.11 Effect of engine load on the percentage recovery of selected PAC (a) and nalkanes (b) at 3000 rpm.
l43
a)
"1::1 120 • E
0 ~ 100 "' c:
0 (J
• 80 ~ -- 60 0
Cl
0~ ...... () 40 :I: ::> o-......_ - 20 0 0---Cl 0
E 0 0 20 40 60 80 100
Percentage of Full Load
b) 8 0
"1::1
~c • E ~
"' c: 6 0
~0 (J .. ~ -- 4 ~0 0
Cl
~c --CJ 2 .§
(I) w .._
0 0 20 40 60 80 100
Percentage of Full Load
c)
6
5 0
0 z 4
11 ....: 0 E (I)
3
.r:. (J
2 / 0 "' 0
ID
~0 --0
0 0 20 40 60 80 100
Percentage of Full Load
Figure 4.12 Effect of engine load on the emissions of unburnt hydrocarbons (UHC)(a), TESSA extracted sample (TES) (b), and bosch smoke (c) at 3000 rpm.
144
At low speed, the associated low swirl has little effect on the fuel sprays, except to carry the
fuel vapour from the spray boundaries beyond the limit of combustion (Rao et al. 1993). At
mid-speed the optimum swirl dynamics and timing enable efficient combustion to develop,
resulting in the maximum power band for small high-speed DI engines. At high engine speeds,
high swirl causes the spray tip to become unstable, resulting in wall jet break down and over
mixing (Rao et al. 1993). Over swirl combined with the limited time for combustion, leads
to increased SOF emissions (Shiozaki et al. 1980, Jensen & Hites 1983, and Ziejewski et al.
1991).
As engine speed is increased, cycle time decreases, leading to increased wall temperature and
reducing time for heat loss; thus increasing air temperature (Campbell et al. 1981 and Li
1982). Bechtold et al. (1984) found that the exhaust temperatures for a 5.7 I eight cylinder
diesel increased from 222°C at 30 mph to 325oc at 55 mph. The overall effect of increased
engine speed is to shorten ignition delays, increase reaction rates, and reduce time for reaction
to occur (Campbell et al. 1981).
The differences between the emissions at low and high speeds may be explained in terms of
varying rates of combustion (kinetics of the reactions) for the compounds. Hence, increased
speed limits the combustion process reactions allowing some organic molecules to react whilst
allowing insufficient time for others to burn to the same extent. In this way, a wider range of
combustion efficiencies may occur. Alternatively, the increased range of recoveries may
indicate that some compounds are produced as a consequence of specific
pyrosynthetic/pyrolysis reactions. The results obtained do not allow survival derived from
varying combustion efficiencies to be distinguished from pyrosynthetic/pyrolysis contributions.
4.2.4 Combined Effect of Engine Speed and Load on Primary Emissions
Figure 4.13 shows the combined effect of speed and load on the recovery of PAC. This clearly
shows the area of greatest emission variability is associated with low load for the three speeds.
Figure 4.13 indicates that for low load there is an optimum speed of 2000 rpm at which
combustion efficiencies are greatest.
145
>-... ., > 0 u ., a: ., 0 .. -c • u ... • Q.
>....
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.20
1.00
: 0.80 0 u ., a: ., 0 .. -c
0.60
., 0.40 u ... .,
Q.
0.20
0.00
a)
0
0
c )
0
1>..
0
a ······· ... '*- ', ~a···· ···- .. ··.o_----o - ·· ··· -~ - . '
20 40 60 80 100
Percentage of Full Load
· ..
20 40 60 80 100
Percentage of Full Load
>-.... • > 0 u • a: • lOt
• -c • u .... • Q.
>.... • > 0 u • a: • lOt • -c • u .... • Cl...
1.2 b)
1.0
0
0.8
0.6 .. 0.4
0.2
0.0
0
1.2 d) 0
1.0
0.8
0.6
0.4
1000 rpm
2000 rpm
- - - - 3000 rpm
, ..._ o ··.. , , :j ~.--.. ~-~
20 40 60 80 100
Percentage of Full Load
... ""· · ~- 0
···a . . . o--~+··w··- -....
0.2
0.0
0 20 40 60 80 100
Percentage of Full Load
Figure 4.13 Effect of engine speed and load on the percentage recovery of fluorene (a), dibenzothiophene (b) , phenanthrene (c) , and pyrene (d).
146
Table 4. 7 also correlates mid-speed with optimum combustion, as indicated by the smallest
decrease in percentage recoveries, when increasing load. At higher loads the difference
between different engine speeds diminishes.
Table 4.7 Effect of Speed on the Percentage Recovery (PR) of PAC between 1% and 25% of Full Load
Speed Difference in PR Difference in PR of Difference in PR of Difference in PR of
(rpm) of fluorene at 1 % dibenzothiopbene at phenanthrene at 1 % pyrene at 1 % and and 25 % of full I % and 25 % of full and 25 % of full 25 % of full Load Load Load Load
1000 0 .65 0.66 0.66 1.02
2000 0 .04 0.08 0.11 0.16
3000 0.18 0.14 0.19 0.10
Similar trends with engine speed and load were found for selected n-alkanes (Figure 4.14 and
Table 4.8). The difference between the combustion efficiency at low speed and low loads was
not as great as for the PAC.
Table 4. 8 Effect of Speed on the Percentage Recovery (PR) of n-alkanes between 1% and 25% of Full Load
Speed Difference in PR Difference in PR of Difference in PR of Difference in PR of (rpm) of C14 at 1% and C16 at 1% and 25% Cl8 at 1% and 25 % C20 at 1% and
25 % of full Load of full Load of full Load 25 % of full Load
1000 0.38 0.29 0.35 0.38
2000 0.06 0.06 0.05 0.04
3000 0.12 0.07 0.06 0.07
The greatest decline in the percentage recoveries induced by increasing the load at low speed ,
indicates that the low swirl and low temperature conditions at low engine power result in the
lowest combustion efficiencies. The results agree with the those of Bazari & French 1993 who
found a significant reduction in the UHC emissions could be achieved by increasing the swirl
at low loads. At high loads the authors found swirl led to over mixing and increased UHC
emissions.
147
With regards to the Prima exhaust, higher speeds resulted in the greatest combustion
efficiencies at full load, indicating that the higher temperatures associated with elevated speeds
may have countered any over-mixing effects. The increase in emissions at low load and high
speed compared with those at mid-speed, may result from increased bulking quenching of the
combustion process due to the volume expansion occurring in the expansion stroke (Yu et al.
1980).
The majority of previous investigations into selected organic emissions have concentrated on
the effect of engine load. However, Zeijewski et al. 1991 also found the highest combustion
efficiencies for PAH were associated with a mid-speed of 2200 rpm.
4 2.5 Comparison of the n-A!kanes with the Mllior PAC emjssjons
Comparing the aromatic with then-alkane emissions (Tables 4.3 & 4.4), reveals that for all
exhaust samples, the combustion efficiencies for n-alkanes are greater than those for PAC. For
example, the comparison of similar boiling point compounds, such as fluorene with
hexadecane and phenanthrene with nonadecane, shows higher percentage recoveries for the
PAH at all engine conditions (Figures 4.15 & 4.16).
It is important to note that unlike the aromatics, the n-alkane emissions have not been corrected
with internal standards. However, spiking of solvent mixes with n-alkane standards and then
working up the mixture as for TES showed that laboratory losses for the n-alkanes with a
higher relative molecular mass than Cl3 were minimal (Section 3.2.2.7). For example, the
laboratory losses for Cl4 and Cl6 were ea. 22% and ea. 10% respectively. Comparing the
n-alkane laboratory losses with the averaged loss of ea. 14 % of the aromatic internal standards
shows that the difference in emission levels is not a consequence of laboratory losses.
This may indicate that at the temperatures and pressures associated with diesel combustion, the
aromatic structures survive relative to the n-alkane structures. The other possibility is that
pyrosynthetic reactions favour the production of the aromatics compared with the n-alkanes.
The study by Abbass (199la) on a 4 1 DI engine using dilution tunnel and filter collection, also
found higher survivals for the P AC in comparison with the n-alkanes.
149
a)
1.0
0
>- 0.8 .... • > 0 u • 0.6 a: • Q Ill 0 •• "' -c
~D----D • u .... • 0.2 0 0
Q..
"' "' "' "' 0.0
0 20 40 60 80 100
Percentage of Full Load
Figure 4.15 Comparison of the percentage recovery of fluorene compared with hexadecane at lOOO rpm (a), 2000 rpm (b) , and 3000 rpm (c).
150
., 1.0
c > 0.8 .... • > 0 u • 0.6 a: • at A • 0 • .&
~-.. c • u .... ---0 • 0.2 0.. -o 0
A A A A
0.0 0 20 .&0 110 80 100
Percentage of Full Load
b)
1.0 c Phenanthrene
A Nono~decane
> 0.8 .... • > 0 u • 0.6 a: • at
"' 0 . .& -c • 0----u .... • 0.2 0.. 0 0 c 0
A--A A A A
0.0 0 20 .&0 60 80 100
Percentage of Full load
cl
1.0
> 0.8 .... • > 0 0 • 0.6 a: Cl at
0~ "' 0.4 -c Cl 0 .... 0
0---• 0.2 0.. 0 0 A---A A A--A
0.0 0 20 40 60 80 100
Percentage of Full load
Figure 4.16 Comparison of the percentage recovery of phenanthrene compared with nonadecane at 1000 rpm (a), 2000 rpm (b), and 3000 rpm (c).
151
However, the authors, due to exhaust deposits and oil survival, were unable to attribute the
greater PAC emissions to pyrosynthesis. The pyrosynthetic possibility is studied further in
section 4.2.6.
The highest percentage recovery for aromatics were obtained for the lighter PAC, especially
at high speed and low loads (Figure 4.17a). The results agree with the concurrent research of
Tancell 1995 at the University of Plymouth. Tancell proposes a link between the reactivity of
the PAC and the survival of the compound, such that the greater reactivity of pyrene compared
to fluorene results in a greater survival of fluorene. The n-alkane percentage recovery also
decreased with increased molecular weight at low load and high speed (Figure 4.17b). As the
chain length of the n-alkanes is increased the percentage recovery decrease, showing that fuel
survival unchanged cannot account for all the n-alkane emissions. If fuel survival unchanged
was solely responsible for then-alkane emissions at low load and high speed, the percentage
recovery would be similar for all n-alkanes, i.e. similar exhaust to fuel ratios. This is indeed
the case at high load and high speed, showing once again that the combustion environment at
this condition supports fuel survival unchanged.
4.2.6 The contribution of combustion generation to the organic emissions
The main source of the major exhaust organics has been shown to be from fuel survival, with
minimal oil contributions. However, as explained in the previous sections the wide variation
in the percentage recovery at low loads may indicate combustion chamber reactions, such as
pyrosynthesis, involved with both n-alkane and aromatic structures.
The combustion chamber generation contribution to the aromatic emissions was further
investigated by comparing the percentage recovery of unsubstituted PAC with the methyl-PAC
derivatives. Figure 4.18a compares the percentage recovery of dibenzothiophene with three
methyl-dibenzothiophene isomers at low engine speed. The numbering of the isomers is purely
arbitrary and does not refer to the position of the methyl- moiety. There are very close
percentage recoveries for both unsubstituted PAC and methyl-PAC. The close ratios at all
loads suggest that dibenzothiophene and the methyl derivatives survive combustion unchanged,
providing further evidence for fuel survival unchanged at 1000 rpm.
152
>-... C)
> 0 0 C)
a: C) Q ftl -c C) 0 L.. C)
a..
0.75
0.60
0.45
0.30
0.15
a)
c 1X of Full Load 4 25X of Full Load v SOX of Full Load <> 75X of Full Load * 1 OOX of Full Load
a
0.00 ..l...r-----.....,...-----------,
0.75
>- 0.60 L.. C)
> 0 0 C) 0.45 a:
C) 0)
!! 0.30 c C) 0 L..
~ 0.15
166
b)
o-o
178
Molecular Weight
c 1 X of full load 4 25X of full load V SQX of full load <> 75X of full load * 100% of full load
202
~ *--=·~ 0---o-o~g~ o----o
-v=--=,..,• *-*-- *-----*- - ---* 0.00 ..l...r----r--.....,...------r------~
14 15 16 18 20
Carbon No.
Figure 4.17 Effect of molecular weight on the percentage recovery of PAH (a) and nalkanes (b).
153
•I 1.0
o.a
~ • .. 0 ... o.a • a: • .. • c OA • ... • A.
0.2 =t=a;;:;:=: 111
0.0 0 20 ~ 10 10 100
Perc.ntage of Full Load
b) 1.0
0.8 4 Oibenzothiophene ... o Me1hyi-Olbenzothlophenej1 I .. o Methyi-Oibenzothiophene 2) • * Methyi-Oibenzothiophene(J) > 0 u 0.1 • a:
• A .. ~ !! 0.4 c: • ft~A---:! • ~~A~ CL
0.2 ~~~A-A ~......_.
0.0 0 20 40 60 10 100
Percentage of Full Load
c l 1.0
<> Fluorene o.a 4 Methyl-Fluorene PI
~ o Methyl-Fluorene 2
• o !.~e thyl-Fluorene (3) >
0 * Methyl-Fluorene (4) 0 u o.s
~0 • a: • .. ~~ !! OA c •
~ • 0~:~ CL 0.2
A i~l 0,0
0 20 40 60 10 100
Perc.ntage of Full Load
Figure 4. 18 Comparison of dibenzothiophene with methyl-dibenzothiophenes at 1000 rpm (a) and 3000 rpm (b). Comparison of fluorene with methyl-derivatives at 3000 rpm (c).
154
Figure 4.18b shows the same comparison but for the higher engine speed of 3000 rpm. The
recovery of dibenzothiophene is consistently higher than the methyl derivatives at each load
sampled. The greatest recovery difference was 0.14% at the lowest load, decreasing to 0.075%
at full load. This indicates that the combustion environment at 3000 rpm may cause
dibenzothiophene to survive to a greater extent relative to its methyl derivatives, especially at
low loads. Alternatively, pyrosynthetic/pyrolytic reactions resulting in the dibenzothiophene
recovery being higher than that of the methyl derivatives may be favoured at low loads and
high speed. Figure 4.18c indicates that a similar process may be involved with fluorene at high
speed. Again, the recovery of the methyl derivatives at low load are much less than the
recovery of fluorene. The difference in recovery decreases as the load is increased, resulting
in near equal recoveries for fluorene and its methyl derivatives at high loads. This further
reinforces the suggestion that conditions of high speed and low load give rise to variations in
P AC recoveries.
In a similar study using GC/MS operated in El mode, the concentrations of the unsubstituted
parent PAC compound to the alkyl derivatives was described by Trier et al. 1990 for a light
duty IDI engine. The study using TESSA showed that in contrast to the fuel, the PAC
concentration in the exhaust samples were greater than the alkyl derivatives, indicating
pyrosynthetic reactions may have been involved.
The study of the recovery of PAH in the exhaust from DI and IDI engines operated on a
variety of fuel blends by Shore (1986) found evidence for pyrosynthetic reactions occurring
for fuels virtually free of PAH. The pyrosynthetic activity was evident from the large
percentage recovery of PAH, for example 2351% of benz(a)anthracene was recovered in the
exhaust of the DI engine. Clearly, the source of the PAH emissions was derived from sources
other than fuel survival.
There are three possible combustion generation routes which could account for the observed
recovery of the PAC relative to corresponding methyl derivatives. Firstly, the methyl- group
may have been cleaved resulting in the formation of the parent PAC. Secondly, the parent
PAC may be derived from aromatic species other than the methyl derivative, and thirdly, the
parent PAC may be derived from aliphatic precursors.
155
Recent radiolabelled research has shown that at 2500 rpm and mid-load, proportions of 1-
methylnaphthalene and 2-methylnaphthalene are converted into naphthalene during combustion
(Pemberton 1995). Further radiolabelled experiments with larger ring PAC methyl derivatives,
such as methyl-fluorene would be needed to establish whether methyl group cleavage occurs
for other PAC. Spiking of fuels with ethyl-phenanthrene and butyl-naphthalene showed that
the formation of vinyl-PAC were favoured for combustion of PAH with larger alkyl side
chains (Pemberton 1995 and Tancell et all995b). In the ethyl-phenanthrene experiment, the
formation of the vinyl-phenanthrenes occurred at 3000 rpm and low load, an engine region
identified as a possible pyrosynthetic/pyrolytic area by this study.
The combustion generation reactions involving the n-alkanes were examined by visually
comparing the distributions of the n-alkanes in the fuel with the corresponding distribution in
the aliphatic fraction derived from high speed and low load (Figure 4.19). The same series of
n-alkanes are present in both samples, but the maximum n-alkane component has shifted from
C16 in the fuel to C15 in TES. Compared with the fuel, the lower carbon numbers in the
exhaust are also relatively higher compared with the higher molecular weight n-alkanes. For
example, the abundance of C14 is greater than C17 in TES, whereas in the fuel Cl4 is lower
in abundance than C19. The skew to the lower volatile carbon numbers, cannot be accounted
for by laboratory losses. Laboratory losses would tend to have the opposite effect, and
decrease the abundance of the lower more volatile molecular n-alkanes relative to the more
stable higher molecular n-alkanes. This suggests that the lighter n-alkanes are to some extent
being derived from pyrosynthetic/pyrolytic reactions, or are less prone to destruction. This
finding contrasts the study by Williams et al. (1989) on a medium-duty 4 I Perkins DI engine,
for which the higher relative molecular mass n-alkanes were shown to survive to a greater
extent than the lighter n-alkanes. The authors propose that this indicates a relatively greater
difficulty for complete combustion of the larger n-alkane structures. The differences between
the two studies may be a consequence of the two differing engine sampling systems used. The
TESSA used in this study has been to shown to collect more of the volatile gas components
of diesel exhaust compared to the dilution tunnel/filter collection used by Williams et al.
(1989) (Trier et al. 1988). Hence, the lower survival attributed to the more volatile n-alkanes
may be a result of a relatively lower collection efficiencies with respect to the larger less
volatile n-alkanes.
156
100 -
I I I, j J
I
100
100
100
Cl6
~ ~ ' J..lo .l. 11 I I
I
200
Temperature (°C)
Cl6
200
Temperature (°C)
(a) Fuel
I
300
(b) TES at
Low Load &
High Speed
300
Figure 4. 19 Comparison of the n-alkane distribution in diesel fuel (a) with the TESSA extracted sample (TES) at 1% of full load & 3000 rpm (b).
157
The TESSA arrangement also precludes post-combustion reactions to a large extent, and in so
doing, quenchs further reactions of the smaller carbon units that may normally occur in the
exhaust pipe. Within the dilution tunnel assembly, there would be a greater opportunity for
reaction and removal of such species. The other possibility is that the engine and the conditions
at which the engines are sampled at, are critical to then-alkane profiles. Indeed, in this study
the profile of selected n-alkanes at 3000 rpm and high load is different to that at the same
speed and low load. At the high load, then-alkane recoveries are much closer compared with
at the low load, and the skew to more volatile n-alkanes found at low load is not present at the
high load.
The ratios of selected n-alkanes relative to hexadecane for fuel and exhaust samples are given
in Table 4.9. For low speed, the normalised ratios for Cl8 and C20 in the exhaust are close
to those in the fuel, whereas those for Cl4 and Cl5 are higher in exhaust. At mid-speed, the
lighter n-alkanes compared with C 16 increase in the exhaust whereas the heavy n-alkanes
decrease with respect to C16. At high speed, the greatest difference between the fuel and
exhaust are apparent at low loads. As the load is increased the ratios in the fuel and exhaust
become closer, confirming earlier findings that the emissions at high load are derived from fuel
survival unchanged.
The skewing effect of the carbon number distribution was also found by Nelson 1989 for the
emissions from a 3 I IDI diesel truck engine. The maximum n-alkane in the exhaust
distribution was C14 compared with Cl5- Cl7 in the fuel. Nelson suggested that gas phase
cracking reactions of the higher n-alkanes may have occurred. In the study significant
quantities of styrene, indene, and naphthalene were found in the exhaust. These compounds
were not present in the fuel. Nelson suggests that a significant proportion of the one- to three
ring aromatics are formed by pyrosynthetic addition of aromatic radicals to unsaturated
aliphatics. Other studies have indicated that simple aliphatic fuels lead to the combustion
formation of PAC (Crittenden & Long 1973, Cole et al. 1984, Henderson et al. 1984, Abbass
et al. 1988, and Barbella et al. 1989). However, synthetic fuels are not accurate
representatives, since the_ complex combustion reactions which occur with standard diesel fuel
would be distorted if a simplex mixture was used.
158
Table 4.9 n-Alkane areas normalised to Hexadecane for a range of Speeds and Loads
Cl4/Cl6 C15/Cl6 Cl8/C l6 C20/Cl6
Fuel 0.57 0.82 0.68 0.49
Speed Percentage of FuU Load (rpm)
1000 1 0.80 1.06 0.75 0.58
25 0.95 1.14 0.59 0.45
50 0.82 1.09 0.63 0.43
75 0.86 1.13 0.65 0.41
100 0.80 1.08 0.70 0.49
2000 1 0.72 1.22 0.51 0.30
25 0.87 0.59 0.45 0.26
50 0.88 1.23 0.46 0.28
75 0.86 1.18 0.46 0.24
100 0.76 1.25 0.48 0.24
3000 1 0.96 1.31 0.51 0.35
25 0.97 1.31 0.45 0.27
50 0.81 1.18 0.43 0 .22
75 0.67 1.07 0.52 0.25
100 0.47 0.87 0.66 0.35
Tancell (1994) used 14C-hexadecane to investigate the pyrosynthetic contribution to the n
alkane emissions for the Prima engine at 2500 rpm and mid-load. He showed that a significant
proportion of the hexadecane (66 %) in the emissions was derived from combustion
generation. Tancell also found that the n-alkane distribution in the exhaust was skewed in
favour of the lower relative molecular mass n-alkanes compared with the fuel distribution,
confirming the findings from this study at a particular engine condition. This may, as Nelson
(1989) proposed, indicate pyrolytic cracking of aliphatic compounds into smaller units.
159
4.3 Comparison of Primary Emissions collected using Initial & Upgraded TESSA This section examines the effect of engine conditions on the n-alkane and PAC emissions
obtained using the upgraded TESSA (Sections 3.3 & 3.4) and compares the results to those
found from the first sampling session using the initial TESSA (Section 3.1). Section 4.3.1
examines more closely the effect of low loads on the combustion of fluorene at low engine
speed. The effect of engine load at 1500 rpm, 2500 rpm, and 3500 rpm for selected n-alkanes
and PAH is presented in Section 4.3.2. Finally, the results from the sampling sets obtained
with the initial and upgraded TESSA configurations are combined at low load in Section 4.3.3.
4.3.1 Combustion of Fluorene at low loads and low speed
The greatest percentage recoveries, and hence lowest combustion efficiencies, for both PAC
and n-alkanes were found at low loads from the profiling obtained with initial TESSA
configuration (Section 3.1). For this reason, the initial testing of the upgraded TESSA
examined the effect of low loads more closely. Loads corresponding to 1% (cold and hot
start), 8%, 17%, and 25% were sampled (Section 3.3.2), aromatic fractions isolated by silica
gel clean-up (section 3.4.2.2.2) and fluorene quantified using GC-FID.
Figure 4.20 shows the effect of the incremental increases in engine load at 1000 rpm, and also
compares the fmdings to those obtained from the first sampling session. The 0.91% recovery
of fluorene at 1 % of full load from the initial TESSA sampling compared with the upgraded
TESSA result (0.96%) is good, as is the 0.26% and 0.28% recovery at 25% of full load for
the initial and upgraded systems respectively. This shows the reproducibility of both the engine
combustion and the sampling system on different days. Increasing the engine load from 1%
to 8% of full load resulted in a much greater combustion efficiency of fluorene. Hence, the
large increase in combustion efficiency observed when the load was increased from 1 % to 25%
is in fact due to ea. the first 30% of the rise. Since temperature was identified as a key factor
in determining the emissions, one of the samples collected at 1% of full load was taken without
conditioning, ie. cold start (conditioning consists of 1 hour at 3000 rpm and 90 Nm, followed
by 15 minutes at the chosen engine condition). As indicated the lack of conditioning increased
the percentage recovery of fluorene from 0.96% to 1.31%.
160
1.50
>- 1.25 .... G) > 0 1.00 () G)
a: G) 0.75 t7)
<V -c: 0 .50 G)
() .... G)
a.. 0.25
0.00 0
' ' ' ' ' ' ' 0 '
v Cold Start
Upgraded TESSA
Initial TESSA
0...- "-8------ --- o ----- -- -- o --- - --- - -0
20 40 60 80 100
Percentage of Full Load
Figure 4.20 Effect of low loads on the percentage recovery of fluorene.
161
Kraft & Lies (1981) also found higher PAH emissions from light-duty diesel engines from the
cold start FfP transient test cycle compared to the hot start for the same cycle. The degree to
which fuel is mixed beyond the lean flammability limit should be the same for both
conditioned and unconditioned engines, since the air:fuel ratio remains the same. Thus, the
decrease in the combustion efficiency provides further proof that low combustion chamber
temperatures causes greater fuel survival.
4.3.2 Effect of Engine Load on the Percentage Recovery of Primary Emissions
The same negative relationship of engine load with percentage recoveries and the associated
increased combustion efficiencies as found for 1000 rpm, 2000 rpm, and 3000 rpm was
replicated by 1500 rpm, 2500 rpm, and 3500 rpm (Figures 4.21 & 4.22). The lower laboratory
losses incurred for naphthalene (calculated from d8-naphthalene) enabled the percentage
recoveries of naphthalene to be calculated. The laboratory losses were still high, owing to the
increased hexane volume used in the revised silica gel clean-up (Section 3.4.2.2.2), removing
a large proportion of naphthalene into the aliphatic fraction. The effect of load on the recovery
of naphthalene was the same as for other PAC at 1500 rpm and 3500 rpm, whilst at 2500 rpm
the recovery was markedly increased at mid-loads compared with other PAC. For all PAC,
the reproducibility of the percentage recoveries are generally good, with a few exceptions.
Fuel surv1vmg unchanged remains the dominant process at high loads, shown by the
convergence of the percentage recoveries, most noticeably for the n-alkanes at 3500 rpm
(Figure 4.22b). Of interest is the distinct increase in the percentage recovery of pyrene at high
load and 3500 rpm relative to the other PAC (Figure 4.2lc). The fact that the remaining PAC
converge at high load whilst pyrene increases suggest that a proportion of the emitted pyrene
was derived from pyrosynthesis, rather than as a result of differing combustion efficiencies.
Tancell (1995) using radiolabelled experiments on the Prima engine, found that pyrene was
especially prone to pyrosynthesis compared to other PAC. For example, Tancell calculated that
of the pyrene recovered in the exhaust, 70% was pyrosynthesized compared with less than
20% for benz(a)pyrene under the same engine conditions.
162
a)
1.2
>- 1.0 .. G > 0 0.8 u Cl a: Cl 0.6 t7l
"' -c 0.4 G -9--u ..
~~ Cl 0.. 0.2 ±
i i i
0.0 0 15 30 45 60 75 90
Percentage of Full Load
b)
1.2 * Naphthalene c Fluorene
>- 1.0 A Dibenzothiophene ... v Phenanthrene Ill > <> Pyrene 0 0.8 0 Ill a:
~----* 0 0.6 t7l
~~i~, "' -c
0.4 G 0 .. G
0.. 0.2
0 • 0 .0
0 15 30 45 60 75 90
Percentage of Full Load
cl
1.2
>. 1.0 0 ... Ill > 0 0.8 0 Ill a: 4) 0.6 t7l .., - ~I c
0.4 Cl u .. 4) 1!1
0.. 0.2 =====---- J. ~ I a
0.0 0 15 30 45 60 75 90
Percentage of Full Load
Figure 4.21 Effect of engine load on the percentage recovery of selected PAC at 1500 rpm (a) , 2500 rpm (b) , and 3500 rpm (c).
163
>~
CD > 0 0 CD ~
CD C7J tU -c CD
0.60
0.45
0.30
~ 0.15 CD
a..
a)
D C14
~ C15
V C16
o C18
* C20
T
~=====-------? ~ 0 - -A~ V ~ -~
- V ~---------i ~
0.00 -'-r----,----.------,..-----r---.----..-0 15 30 45 60 75 90
Percentage of full load
bl 0 .60
>-~ 0 .45 CD > 0 0 CD a:
0.30 T
CD
~~~ C7J tU -c CD 0 0.15 ~
CD
~ -------===:~ a.. T
0 .00
0 15 30 45 60 75 90
Percentage of Full Load
Figure 4.22 Effect of engine load on the percentage recovery of selected n-alkanes at 1500 rpm (a) and 3500 rpm (b).
164
Then-alkane recovery follow the PAC trends, with the recovery remaining lower than those
of equivalent boiling point PAC. The highest recoveries are, as before, associated with the
lighter n-alkanes and PAC.
4 3 3 Effect of Engine Spwl for the Combined Sampling Sets
The combination of the sampling sessions at 1% of full load for all speeds on the emissions
of fluorene and hexadecane is shown in Figure 4.23. The data for hexadecaoe at 2500 rpm is
missing due to the ODS cartridge clean-up method being employed. Taking into consideration
the error bars, the recovery of both compounds is highest at low load and low speed. The
greatest combustion efficiencies occur at mid-speeds, with the recovery increasing with further
speeds. These trends confirm the findings from the sampling sessions carried out with the
initial TESSA. Hence, the organics survive at low speeds and low loads due to the limited
temperatures and insufficient air motion, whereas at high speed the combustion is restricted
by the limited reaction time and over-swirl. The recovery of hexadecaoe, whose volatility is
similar to fluorene, is always lower than that of the latter, and as discussed may mean that the
structure of fluorene is more resistant to combustion than that of hexadecane. Alternatively,
the combustion chamber generation of fluorene may be relatively higher than that of
hexadecaoe.
165
1.2
Fluorene 1.0
~ ----- Hexadecane >-..... CV 0.8 > 0 0 CV a:: ~
CV 0.6 ~T C)
«:J ~ ~ u c 1 CV ~ 0 0.4 q 1~~ .....
\ CV \
~~ a_ \ \
\ ..J..
0.2 \ \ ----D-----o
\ -- -tJ- ----a--
0.0 1000 2000 3000 4000
Speed
Figure 4.23 Effect of speed on the percentage recovery of fluorene and hexadecane at low load.
166
4.4 Summary of Major Organic Emissions
1) Fuel survival is dominant over lubricating oil survival for the selected organic compounds
studied.
2) Combustion efficiencies (as defined by percentage recoveries) are lowest at low loads, and
increase with load. The highest emissions at 1% of full load and 1000 rpm are a consequence
of the low temperatures and the fuel:air mixture being beyond the lean flammability limit.
3) Then-alkane emissions produce the same trends as for the PAC emissions, indicating that
the n-alkanes and aromatics are derived primarily from the same source. The selected organic
emissions agree with the overall unbumt hydrocarbon and TES emissions. The dominant
source for such emissions is via fuel survival.
4) The convergence of the percentage recovery at high load indicates that the combustion
environment at these conditions favours fuel survival unchanged. The combustion environment
at high load results in the greatest temperatures with the smallest range or gradient.
5) The poor mixing at low loads results in lower overall temperatures with a larger temperature
gradient than at high load. This may result in a greater range of combustion efficiencies. Low
load combined with high speed may restrict the kinetics of the reaction and result in a wider
range of combustion efficiencies than at lower speeds. Alternatively, the combustion
environment at low loads and high speed (defined by a wider range of temperatures) may
favour greater combustion generation, such as the cleavage of methyl- groups from PAC.
6) The optimum swirl, time and combustion chamber temperatures associated with mid-speed
results in greatest combustion efficiencies at low loads. The higher temperatures at high load
results in speed having a smaller effect at high loads.
167
7) The percentage recovery of the PAC were significantly higher than those of equivalent
boiling point n-alkane molecules. The difference may arise from the P AC structures being
more stable than the straight chain n-alkanes, resulting in greater survival of the PAC.
Alternatively, the pyrosynthetic formation of PAC may be favoured.
8) The emissions of some PAC are more likely to be derived from pyrosynthesis than others.
Pyrene especially, appears atypical in behaviour compared to other PAH and seems to be more
readily pyrosynthesized.
9) The skewing of n-alkane profile suggest combustion generation is involved at specific
engine conditions, possibly resulting from the breakdown of higher molecular compounds into
smaller n-alkane molecules.
168
Chapter 5- Secondary Nitro- & Oxy-PAC Emissions
This chapter examines the effect of engine speed and load on the secondary nitro- and oxy
PAC emissions from the Prima engine. These PAC are formed from conversion of primary
PAH, such as pyrene, during the combustion and exhaust stages. The results presented in this
chapter were derived from sampling the exhaust stream of the Prima as it leaves the
combustion chamber by using TESSA. In this way, the potential of the combustion chamber
to form nitro- and oxy-PAC can be assessed, and the artefact problems associated with the
dilution tunneVfilter sampling systems avoided. The main part of the chapter is focused on the
nitro-PAC, since some members of this PAC class have be found to be highly mutagenic and
carcinogenic to mammalian biological systems.
A comparison of HPLC fractionations in Section 5.1, evaluates the overall chemical
composition of the exhaust as it leaves the chamber, along with the effect of speed and load
on the relative contributions of different chemical classes. The results from the nitro-PAC
profiling are reported and discussed in Section 5.2, as are those regarding oxy-PAC in Section
5.3. The major findings from both nitro- and oxy-PAC profiling are listed in Section 5.4.
5 .1 The Contribution of Different PAC classes to the Emissions
Fractionation of diesel combustion samples provides information on the relevant contributions
of different chemical classes present in the sample. The elution times of specific classes of
compounds are established by analyses of standards, using semi-preparative normal phase
HPLC (Section 3.4.2.3). The contribution of specific PAC classes, such as nitro-PAC, to the
overall composition can then be investigated.
The overall composition of TES under different speeds and loads showed a regular pattern or
profile (Figure 5.1). The first and largest class of PAC to be eluted (0-26 mins.) corresponds
to the non polar PAH, PASH, and alkyl-derivatives. The source for these compounds is
predominately from unburnt fuel surviving combustion (Chapter 4). By contrast, the
mononitro-PAC HPLC fraction (26-33 mins.) is much reduced in concentrations.
169
lOO
e = ~ 11") M -~ ~ Cj
= ~ .Cl
loo Q .,
.Cl <
100
e = ~ 11") M -~ ~ Cj
= ~ .Cl
loo Q .,
.Cl <
100
e = ~ 11") M -~ ~ Cj
= ~ .Cl
loo Q .,
.Cl <
Key A= PAH, PASH & Alkyl derivatives 8 = Mononitro-PAC C = Dinitro-PAC
10 20
A 10 20
30
30
D = Nitro-oxy-PAC E = Polars
~0
D 40
10 20 30 40 Retention Time (minutes)
50
50
50
(a) Low Load
60
(b) Mid Load
60
(c) High Load
60
Figure 5.1 Overall chemical composition of the aromatic fractions of TES at 1500 rpm and a) low load, b) mid load, and c) high load, determined using normal-phase high-performance liquid chromatography (Chromatographic conditions given in Section 3.4.2.3.2).
170
The second large group of PAC eluted was divided into the dinitro-PAC (33-38 mins.) and
nitro-oxy-PAC (38-48 mins.). Oxy-PAC, such as the ketone and aldehyde PAC derivatives
eo-eluted with the dinitro- and nitro-oxy PAC fractions, whereas the more polar carboxylic
acids elute in the final large band (48-63 mins.). Engine load seems to have little effect on the
relevant distribution of chemical classes at low engine speed. A similar regular trend was
repeated for higher speeds.
The regular emission pattern generated for a wide variation of engine conditions suggest
similar combustion reactions had occurred across the load and speed ranges. Hence, a
proportion of the fuel input to the cylinder has been transformed into the secondary nitro- and
oxy-PAC. Due to the complexity of the HPLC fractions, it is necessary to chemically
characterise each HPLC fraction (Sections 5.2 & 5.3). Individual species can then be followed
over varying speeds and loads, and the factors giving rise to their formation investigated.
52 Nitro-PAC Emissions
The high mutagenicity and carcinogenicity of some nitro-PAC (Tokiwa et al. 1981 & 1986,
Rosenkranz & Mermelstein 1983, and Beland et al. 1985) has been central to health concerns
related to diesel emissions {IARC 1989 & QUARG 1993). Research has shown that 1-
nitropyrene, a powerful mutagen, can contribute between 10 to 40% of the total mutagenicity
of diesel particulate extracts (Schuetzle et al. 1981, Nakagawa et al. 1983, Schuetzle & Frazier
1986, and Veigl et al. 1994). The health concerns in relation to nitro-PAC emissions, makes
elucidation of the engine processes controlling their formation highly desirable.
The formation of nitro-PAC result from the nitration of PAH, such as pyrene. The nitrating
species, termed NO., are generated from the thermal decomposition of nitrogen and air in the
combustion chamber (Scheepers & Bos 1992b). The NO, may then react with PAH to form
nitro-PAC via free radical processes (Nielsen et al. 1983, and Scheepers & Bos 1992b).
Nitration may also occur by electrophilic substitution of the PAH ring by nitrogen dioxide,
N02, in the presence of nitric acid via the nitronium ion (Nielsen 1984, Ross et al. 1988, and
Scheepers & Bos 1992b). The nitro-PAC emissions are limited by competitive removal of
PAH by combustion, in which case nitration cannot occur. Similarly, nitro-PAC may react
171
further, for example further oxygenation may give rise to nitro-oxy-PAC (Schuetzle & Perez
1983). Research to date has provided little information on which nitration reactions are
favoured under different engine conditions.
This section details the emissions of nitro-PAC from the Prima engine and compares the results
with those in the literature. Three types of nitro-PAC are considered, namely, mononitro
PAC, dinitro-PAC, and nitro-oxy-PAC. The aim of the nitro-PAC profiling was to correlate
the abundance of nitro-PAC emissions with the nitration potential of the combustion chamber.
The short transfer line between the engine and TESSA enables the combustion products as they
leave the chamber to be characterised. Post-combustion reactions are significantly reduced for
sample collection using TESSA.
Samples were derived from three speeds and at three loads for each speed (Section 3.4.1). The
samples were extracted, concentrated, cleaned-up, and fractionated by normal-phase HPLC
(Section 3.4.2). Section 5.2.1 reports and discusses the identification and levels of mononitro
PAC in the fuel, oil, and emission samples collected using TESSA and analyzed by the
selected detection systems. The search for the dinitro-PAC and nitro-oxy-PAC in the Prima
emissions is detailed in Section 5.2.2.
5 2 1 Analysis of the Mononitro-PAC HPLC fractions by Gas Chromatography with Electron Capture Detection The mononitro-PAC from the TES, fuel and oil sump samples were identified (by eo-injections
with a standard mix and by RI) and quantified (external calibration) by GC-ECD (Section
3.4.3.1.4). The GC-ECD system provided the greatest detection, resolution, and access needed
for resolving the complex mononitro-PAC HPLC fractions. Figure 5.2 shows the analysis of
three mononitro-HPLC fractions sampled at 1500 rpm and for three load positions. As for the
overall HPLC fractionation profiles, the chemical distribution of the mononitro-PAC HPLC
fractions were found to be similar for varying loads and speeds.
The GC-ECD analysis revealed 1-nitronaphthalene, 2-methyl-1-nitronaphthalene, 2-
nitronaphthalene, and 1-nitropyrene present in all TES. Figure 5.3 focuses on the temperature
elution zone associated with the nitronaphthalenes for the TES collected at 1500 rpm.
172
100
~ ., = Q c. ., ~ ~ u ~
100
lOO
100
200
200
(a) Low Load
300 Temperature ( °C)
(b) Mid Load
Temperature ( °C)
(c) High Load
300
Temperature ( °C)
Figure 5.2 The analysis of the mononitro-PAC HPLC fractions derived from sampling at 1500 rpm and across the load range, by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.1).
173
100
100
100
Key 1 = 1-Nitronapbthalene 2 = 2-Methyl-1-nitronaphthalene 3 = 2-Nitronapbthalene
160 165 170 Temperature ( OC)
2
160 165 170 Temperature ( OC)
3
160 165 170 Temperature ( OC)
(a) Low Load
175 180
(b) Mid
Load
175 180
175
(c) High Load
180
Figure 5. 3 The profile of the nitronaphthalenes at 1500 rpm and across the load range, determined using gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4. 1).
l74
The regular pattern of the nitronaphthalenes is replicated at each load position. The regularity
of the profiles suggests that a common formation pathway for nitro-PAC is operating.
Figure 5.4a shows the analysis of the mononitro-PAC HPLC fraction of the fuel by GC-ECD.
There are very few positive peaks, and during the middle section of the GC temperature
programme some non-electrophilic compounds are evident, as indicated by the negative peaks
(between 30 and 35 minutes). Figure 5.4b&c shows the delimited retention windows of the
fuel analysis, for the retention times associated with mononitro-PAC standards. Since there are
no peaks evident between 22-23.5 minutes (nitronaphthalene standards elution window, Section
3.4.3.1.4.1) nor between 42.0-42.3 minutes (1-nitropyrene retention time, Section
3.4.3.1.4.1), there are no mononitro-PAC in the fuel. Figure 5.5a shows the analysis of the
mononitro-PAC HPLC fraction of sump oil. There are a number of prominent peaks,
however, as for the fuel no mononitro-PAC could be found (Figure 5.5b&c). The absence of
nitro-PAC in the fuel and oil was verified by eo-injection techniques.
The lack of mononitro-PAC in the fuel and sump oil showed that specific nitro-PAC were
formed as a result of combustion reactions, rather than as a consequence of nitro-PAC
surviving combustion. Previous studies by Jensen et al. (1986) found that sump oil
accumulated 1-nitropyrene to a small degree, but could find no evidence for oil survival
contributions to nitro-PAC emissions. The results from this study were expressed as the weight
of nitro-PAC emitted relative to the total TES weight collected (Table 5.1). The laboratory
losses were found to be minimum for the engine sample work-up and fractionation procedures,
for example ea. lO % of 1-nitronaphthalene was lost (Section 3.4.2.5).
175
I 1.0
21
41
20 30 40 Retention Time (minutes)
22 23 Retention Time (minutes)
42 43 Retention Time (minutes)
(a) Overall Fuel
50 60
24
44
(b) Search for Nitronaphthalenes
25
(c) Search for 1-nitropyrene
45
Figure 5.4 The analysis of the mononitro-PAC HPLC fraction of diesel fuel by gas chromatography with electron capture detection (a), and the verification of the lack of nitroPAC in the fuel (b&c). Retention times for nitro-PAC standards: 1-nitronaphthalene (22.237 mins.), 2-methyl-1-nitronaphthalene (22.919 mins.), 2-nitronaphthalene (23 .487 mins.), and 1-nitropyrene (42.054 mins.) (Chromatographic conditions given in Section 3.4.3.1.4.1).
176
I
10
21
41
20 30 40
Retention Time (minutes)
22 23 Retention Time (minutes)
42 43 Retention Time (minutes)
(a) Overall Sump Oil
50 60
(b) Search for Nitronaphthalenes
24 25
44
(c) Search for 1-nitropyrene
45
Figure 5.5 The analysis of the mononitro-PAC HPLC fraction of sump oil (after 50 hours use) by gas chromatography with electron capture detection (a), and the verification of the lack of nitro-PAC in the fuel (b&c). Retention times for nitro-PAC standards: 1-nitronaphthalene (22.237 mins.), 2-methyl-1-nitronaphthalene (22.919 mins.) , 2-nitronaphthalene (23.487 mins.) , and 1-nitropyrene (42.054 mins.) (Chromatographic conditions given in Section 3.4.3.1.4.1).
177
Table 5.1 Nitro-PAC emitted from the Prima Engine at different engine conditions
Speed % of 1-Nitropyrene 1-N itronaphthalene 2-N itronaphthalene 2-Methyl-1-(rpm) Full inTES inTES inTES nitronaphthalene
Load in TES
(ppm) _fup_ml .. CPrml wm~
1 1.6 9.3 7.1 8.2 1500 50 1.1 11.9 6.1 10.2
85 1.3 10.6 7.5 7.7
1 1.0 - - -2500 50 2.6 - - -
85 2.2 - - -
1 0.6 13.3 7.0 7.8 3500 29 0.6 20.9 7 .9 11.7
90 5.3 35.1 18.9 18.5
Notes: 1) Averaged results from duplicate samples 2) Results expressed as 11g of nitro-PAC emitted/g of TES collected
3) Quantification of nitronaphthalenes at 2500 rpm not possible as a result of negative peaks, from aliphatic compounds, interfering with the baseline position.
5.2, 1.2 Analysis of the Mononjtro-PAC HPLC fraction derived from High Speed and High Load by Gas Chromatography/Mass Spectrometry operated in Negative Ion Chemical Ionisation Mode. Analysis of nitro-PAC by GC/MS operated in the NICI mode results in very little
fragmentation (Newton et al. 1982, Oehme et al. 1982, Ramdah1 & Urdal 1982, & Bayona
et al. 1988), and for this reason, selective ion monitoring (SIM) of the molecular ion of nitro
PAC was utilized in this study. The SIM mode greatly enhanced the sensitivity, resulting in
the same order of sensitivity as that of GC-ECD (Section 3.4.3.1.5). Matching of peaks on a
molecular ion basis and retention indices provides strong evidence for the presence of nitro
PAC, since the compound under question not only elutes at the precise time of the proposed
oitro-PAC but also the main ion for the compound matches that of the nitro-PAC standard.
Selective ion monitoring combined with a retention index based on 1-nitropyrene confirmed
the GC-ECD identifications of the nitronaphthalenes. Based on the comparison of the areas of
1-nitropyrene in the sample with that in the standard, the concentration of 1-nitropyrene in
TES at high speed and high load was calculated to be 1. 78 ppm. Considering the crude
GC/MS NICI quantification and that the analysis was performed several months later than the
178
GC-ECD analysis, this figure agrees favourably with that of 5. 75 ppm obtained using GC
ECD.
The GC/MS NICI was then used to search for other nitro-PAC in the sample. This was
achieved by calculating the molecular ions for the nitrated form of major PAH and alkyl
derivatives. Scanning for mono-, di-, and trimethylnitronaphthalenes revealed a series of
peaks, similar to the distribution of naphthalene and methy1naphthalenes in the fuel (Figure
5.6). The spread of the retention times for the methylnitro-PAC is greater than the methyl
PAR due to the greater isomer possibilities.
Similarly, scanning for the molecular ions of nitrofluorene (211) and nitroanthracene (223) as
well as the methylnitro- derivatives indicated a number of compounds present (Figures 5. 7 &
5.8). Investigating the peaks corresponding to ions of molecular weight 211 (by eo-injection),
showed no presence of 2-nitrofluorene. The GC-ECD analysis could also find no evidence for
2-nitrofluorene in the exhaust. This is interesting, considering the levels of fluorene present
in the fuel (Section 4.1). The 9-nitroanthracene isomer was not present among the 223 ions.
However, it is phenanthrene rather than anthracene which is present at elevated levels in the
fuel (Section 4.1). This suggests that the 223 and 237 ion distributions may refer to
nitrophenanthrenes and methy1-nitrophenanthrenes. Even though the exact structures of the
scanned molecular ions have not been elucidated by GC/MS, the analysis has shown a
similarity between the fuel distribution and the secondary emission profiles, ie. a proportion
of the fuel has undergone transformation into nitro-PAC. However, the lack of 2-nitrofluorene
in the emissions suggests that certain elements of the fuel are more susceptible to
transformations than others.
179
~ (.I
= 4"::1 "C = = .c < ~ ~ ·--4"::1 -~ er::
(a) TES at High Speed and High Load
100! .LL . Nitronaphthalenes (173 Ion)
4721 ~. Methyl
nitronaphthalenes (187 Ion)
DimetbylnitronapbtbaJenes
(201 Ion) l92t ;~, 9
'2
1,__...,...-==;=:;::::::::::;=::::::::::;=:::::::~~.--....---.---.-----.~...--::::;:::::T:::::nm;=· .... etbyl
n:nloaphthaleo"
100
100
100
lOO
I 7 -----;
. _ (215 Ion)
16.30 20.10 23.50 27.30
Retention Time (minutes)
(b) Fuel
"'
.N J
.1 ~1 . 0 53 9.11 12.53 16.34 20.1 6
Retention Time (minutes)
N b haJene ap t on) (128 I
Metb ylaJenes on)
napbtb (142 I
Dim et hylaJenes on)
napbth (156 I
Trime tbylaJenes Ion)
naphtb (170
Figure 5. 6 Distribution of the nitronaphthalenes in TES collected at 3500 rpm and high load, determined by gas chromatography/mass spectrometry (GC/MS) operated in the negative ion chemical ionisation mode (a) compared with the distribution of naphthalenes in the fuel, as determined by GC/MS in the electron impact mode (b) (Chromatographic conditions given in Sections 3.4.3.1.5 and 3.2 .4 respectively).
180
100 (a) 211 Ion scanned ~ V
= ~ "0
= = .c < ~ .::: -~ ~ ~
100 ~
(b) 225 Ion scanned V
= ~ "0
= = .c < ~ .::: -~ -~
8.20 16.40 25.00 33.20 41.40
Retention Time (minutes)
Figure 5. 7 Distribution of molecular ions 211 and 225 in the mononitro-PAC HPLC fraction collected at 3500 rpm and high load. Determined using gas chromatography/mass specrrometry operated in the negative ion chemical ionisation mode (Chromatographic conditions given in Section 3.4.3.1.5) .
181
100 (a) 223 Ion scanned
(b) 237 Ion scanned
8.20 16.40 25.00 33.20 41.40
Retention Time (minutes)
Figure 5.8 Distribution of molecular ions 223 and 237 in the mononitro-PAC HPLC fraction collected at 3500 rpm and high load. Determined by gas chromatography/mass spectrometry operated in the negative ion chemical ionisation mode (Chromatographic conditions given in Section 3.4.3.1.5).
182
5.2.1.3 Comparison of the levels and distribution of Nitro-PAC found using TESSA with
Previous Studies
The levels of nitro-PAC found in this study (Table 5.1) are significantly lower than the levels
found by previous studies of light-duty engines collected with dilution tunnel/filter sampling
systems (Table 5.2).
I bl 52 a e p reVIOUS N'tr PAC 1 1 f 1 0- eves rom L' ht D t D ' 1 E IgJ - ury_ 1ese ng_mes
Reference Sample Details 1-N itropyrene
Li & Westerholm 1994 Light-duty diesel engine sampled with cyclone 4.5 JJ.glg of extract particulate separator
Veigl et al. 1994 1.8 I normal aspirated DI diesel operated on transient cycles: HWFET 854 ng/filter FTP 69 ng/filter
Imaizumi et al. 1990 Diesel exhaust 50 ng/g of particulate
Williams et al. 1986 Light-duty Oldsmobile on HWFET (48 mph steady I 07 ppm in extract speed) sampled with dilution tunnel
Brown & Poole 1984 Diesel exhaust sampled with dilution tunnel 165 ± 15 JJ.glg of extract
Schuetzle et al. 1982 Four light-diesel engines sampled using ranged from 55± 11 to dilution/filter sampling 2280± jJ.g/g of extract
CampbeU & Lee 1984 Light-duty diesel sampled with dilution tunnel 43 JJ.g/g of extract
Paputa-Peck et al. 1983 Light-duty diesel sampled with dilution tunnel 75 ± 10 ppm of extract
Tejada et al. 1982 Two light-duty diesels sampled on FTP using JJ.glg of extract: dilution tunnel: 1978 Oldsmobile 100.0 1980 Vol.kswagen Rabbit 58.5
The study by Li & Westerholm (1994) using a cyclone particulate sampling system, found
levels of 1-nitropyrene comparable with this study. There is a large degree of variation in the
levels between different studies and engines. The levels of 1-nitropyrene found in this study
are more in line with some of the 1-nitropyrene emissions from medium- and heavy-duty
engines (Table 5.3).
183
Table 5.3 Previous Nitro-PAC Emissions from Medium- & Heavy-Duty Engines
Source Sample Details 1-Nitropyrene
Draper 1986 Caterpillar 3304 NA sampled using dilution tunnel at 2 f.Lglg of conditions: particulate: high load and moderate speed 5.0 low load and high speed not detected
Bechtold et al. 1984 Oldsmobile 5 .7 L 8 cylinders. Four cycles of Federal Test 81 ppm in Procedure using dilution tunnel sampling particulate
Jin & Rappaport 1983 Dilution tunnel sampling of four heavy duty diesels: ng/mg of extract:
Cummins VTB-903 0.7 International Harvester DT -466 28.0 Volvo TD-100 9.2 Caterpillar 3046 DIT A 142.0
Nakagawa et al. 1983 1970 Isuzu BY 30 Bus 4 I diesel sampled at idling and 1200 70.5 ppm in the rpm at the end of exhaust with condensation trap followed extract by filter collection
Schuetzle & Perez Heavy-<luty engine sampled with dilution tunnel at 2 ppm in extract: 1983 conditions:
idle 28.0 2100 rpm & full load 1.0
Rappaport et al. 1982 Dilution tunnel sampling of 2 diesels: ng/mg of extract:
medium-<luty DSR-46 8 heavy-<luty 1980 Mack 20
Previous studies which relied on dilution and fi lter collection may have included significant
artefact contributions to the nitro-PAC quantifications. Several experimental studies have
investigated the possibility of artefact formation of nitro-PAC during sampling.
Pitts et al. (1978) found that PAH on glass filters were transformed into nitro derivatives.
Tokiwa et al. 1981 exposed pyrene to 10 ppm of N02 and found the formation of 1-
nitropyrene. This level of N02 could be found in diluted exhaust (Gaddo et al. 1984). Gibson
et al. ( 1981) found that increasing the sampling time increased both the 1-nitropyrene
emissions and mutagenicity whereas the pyrene concentrations decreased. Bradow et al. (1982)
concluded that above 5 ppm N02 the dilution tunnel system would generate artefacts.
184
Lindskog et al. (1983) found that exposure of pyrene to 1 ppm N02 on filters caused
degradation. The degradation was greatly enhanced by nitric acid. Risby & Lestz (1983) used
a theoretical model for pyrene based on the synthetic fuel and oil study by Herr et al. (1982).
Risby & Lestz (1983) concluded that the activation energies for surface reactions on filters
were lower than the gas-phase reactions, hence artefacts reactions were favoured. Schuetzle
& Perez (1983) estimated an average of 12.5% of the nitro-PAC in diesel particles collected
in dilution tunnels resulted from artefact formations.
Gaddo et al. (1984) showed that 1-nitropyrene concentrations increased dramatically with
sampling time for collection on fibre glass filters. For example, the 1-nitropyrene
concentration after 5 minutes of sampling was 24 ± 7 p.g/g compared to 74 ± 7 p.g/g after 10
minutes. The authors claimed that after one hour of engine sampling, somewhere between 50
to 90% of the nitro derivatives could been formed as a consequence of the sampling process
itself. Hartung et al. (1984) found that 18% of nitro-PAC were derived from artefact reactions
on the filters used with the dilution tunnel system. The authors also found electrostatic
precipitator sampling to have a greater tendency for artefact formations. Chan & Gibson
(1985) found that increasing the N02 concentration from 2 to 4 ppm in diluted exhaust for a
23 minute FTP sampling cycle resulted in a 60-160% increase in the level of 1-nitropyrene
emitted.
Care should be taken when interpreting some of the early artefact studies, such as the initial
work of Pitts et al. (1978). Grosjean et al. (1983) point out that such studies tended to
employed high flow rates through the filters and spiked the filters with large doping masses
of PAH, such as pyrene. On-road measurement of 1-nitropyrene in the Allegheny tunnel by
Gorse et al. (1983) found significantly lower emissions compared to dilution tunnel studies.
The authors note that the higher acidity of dilution tunnels compared with the more neutral
environment in the road favoured nitration reactions within the dilution tunnels. Similarly, the
higher levels of N02 encountered in dilution tunnels (2-4 ppm) compared with only 0.04-0.16
ppm in the road tunnel may also favour nitration in the confined dilution tunnel arrangement.
However, significant degradation of nitro-PAC may have occurred in the environment,
accounting for the lower nitro-PAC in the road tunnel study (Stark et al. 1985).
185
The concept of the TESSA results in the minimum opportunity for artefact formation. This is
due to the extremely fast sampling times (order of seconds) and the corresponding rapid
removal of the sample from the incoming exhaust. Artefact formation is an inherent problem
with the dilution tunnel design, since the sample is held on the filter whilst gaseous reactive
species are passed through for lengthy periods (23 minutes for FfP cycle). The ftlter provides
a condensation point for nitric acid which greatly enhances the artefact nitration. Sasaki et al.
(1980) found that dilution tunnels progressively convert NO into N02 with sampling time,
creating a greater nitration potential with extending sampling.
One area of concern involved with the work-up of TES samples, was with the separation of
the extracted organics into the aqueous methanol phase and the DCM phase containing the
PAC of interest. This partition, achieved immediately as the solvent mixture flows from the
tower into distilled water, was introduced by Trier (1988) to quench any post-combustion
reactions. Concern was raised that there may be some interaction of PAH in the DCM with
the acidic aqueous methanol, resulting in nitration of PAH. The possibility that nitro-PAC
were formed from the interaction of the acidic aqueous methanol phase and the DCM organic
phase was examined (Section 3.4.2.1). This was achieved by adding an aliquot of a standard
PAH mix (with each PAH weight added equivalent to that present in a TES) to a pre-extracted
aqueous phase from an engine run. The solution was left for 24 hours. Since, the normal
work-up procedure would be to rapidly separate the aqueous methanol and DCM phases (less
than 30 minutes per sample), any nitration that may have occurred would be detectable in the
nitration check. Following the extraction of the PAH from the aqueous phase and isolation of
the fraction corresponding to the nitro-PAC elution window by NP HPLC, the sample was
analyzed by GC-ECD. The analysis, shown in Figure 3.11 (Section 3.4.2.1), could find no
evidence for any formation of nitro-PAC.
If it is assumed that the careful use of filters and dilution sampling techniques result in artefact
contributions to the nitro-PAC emissions of the order of 10 to 30%, there still exists a
significant difference between the highest 1-nitropyrene emissions from the Prima engine (5.3
ppm) sampled with TESSA, compared with the majority of the light-duty engines sampled with
186
dilution tunneVfilter systems. Some of the previous studies on light-duty diesels, such as that
by Campbell & Lee (1984), are closer to this study (assuming 30% artefact derived, the
Campbell & Lee study gives a concentration of 12.9 ppm of 1-nitropyrene in the extract).
More importantly the dilution tunnel samples are typically derived from engine conditions less
extreme than 3500 rpm and high load, conditions which in this study resulted in the highest
1-nitropyrene levels. The average of the remaining 1-nitropyrene emissions from this study
gives an average of 1.4 ppm in the extract. Even after correcting for severe artefact
contributions these 1-nitropyrene emissions are significantly lower than any of the previous
studies on light-duty engines, sampled using dilution tunnels.
The different between the levels found in this study and previous ones may be a reflection of
the type of sampling system used. In this study the samples are derived primarily from the
combustion chamber with the rninium of exhaust contributions, whereas dilution tunnel
systems are designed to simulate exhaust and environmental effects. Hence, the difference in
1-nitropyrene emissions obtained from this study compared with the dilution tunnel studies
may be due to the limited dilution allowed by the engine and TESSA arrangement. This
suggests that increased dilution may increase the nitro-PAC emissions, possibly by increasing
reaction time between PAH and nitrogen dioxide in the diluted exhaust stream. Li &
Westerholm (1994) collected emissions using a particulate cyclone without any additional
exhaust dilution. The levels of 1-nitropyrene emitted were comparable with those in this study,
supporting the theory that increased dilution increases nitro-PAC emissions, possibly as a
consequence of artefact formations.
Another reason for the different emission levels may result from the engine technology
available at the time of the studies. The early engines would not have been as efficient as the
more modern Prima engine. Consequently, the emissions from the older engines may have
been relatively higher than the Prima design. The comparison of 1978 diesel engines to the
later 1980 models for 1-nitropyrene emissions by both Gibson et al. (1981) and Tejada et al.
(1982) did in fact show a marked reduction for the newer engines.
187
Many dilution tunnel studies have found between 60 to 200 different nitro-PAC in the exhaust
extracts, with the most abundant nitro-PAC in most cases being 1-nitropyrene (Schuetzle et
al. 1982, Paputa-Peck et al. 1983, and Williams et al. 1986b). In this study, analysis of the
sample collected at high speed and high load by GC/MS NICI also indicated that many nitro
PAC were present in TES at low levels (Section 5.2.1.2). The GC/MS results indicate that a
proportion of the fuel P AC undergoes nitration to form nitro- derivatives with a similar
distribution to the fuel. A series of studies by Henderson et al. (1982, 1983, & 1984) also
found that a fraction of fuel underwent nitration, resulting in the exhaust nitro-PAC
composition being similar in distribution to the PAH in the fuel.
By contrast to other light-duty diesel engine studies, the nitronaphthalenes were in greater
abundance than 1-nitropyrene. The study of diesel nitro-PAC emissions by Williams et al.
(1986b) found much lower levels of 1-nitronaphthalene (0.3 ppm of extract) compared with
1-nitropyrene (107 ppm of extract). Heavy-duty diesel studies have in some cases found higher
1-nitronaphthalene emissions, for example Draper (1986) could not detect 1-nitropyrene at
high load and moderate speed whilst 1-nitronaphthalene was emitted at 0. 77 p.g/g particulate.
However, at a higher speed and lower load, the situation reversed, with 1-nitropyrene
emissions of 5.0 p.g/g particulate compared with 1-nitronaphthalene emissions of only 0.47
p.g/g particulate. The greater nitronaphthalene levels compared with 1-nitropyrene found by
this study may to a limited extent be a result of higher laboratory losses of ea. 30% for 1-
nitropyrene compared with the nitronaphthalenes (for example ea. 10% of 1-nitronaphthalene
lost). However, taking this into consideration still leaves the nitronaphthalene levels higher
than those of 1-nitropyrene. The different distribution of the nitro-PAC in the Prima emissions
may indicate different nitro-PAC formation processes may have been involved. This may again
reflect the decreased potential for post-combustion reactions, including artefact contributions,
using TESSA. Alternatively, some of the dilutions tunnel studies may not have collected the
more volatile nitronaphthalenes to the same extent as 1-nitropyrene. The PAH content of the
fuel may also have varied significantly between the different studies, and it may be that in this
study there were greater levels of naphthalene compared with pyrene than there were in other
studies, or alternatively other studies had relatively higher pyrene levels.
188
5 2 1.4 The Effect of Engine Speed and Load on the Mononitro-PAC Emissions
The rapid sampling and low artefact potential of TESSA enables the effect of engine speed and
load to be examined. Nitro-PAC profiling over a wide range of speeds and loads is best suited
to TESSA, which outperforms other engine sampling techniques in terms of mass sampling.
Figure 5.9 shows the effect of engine load on the identified nitro-PAC emissions at the lowest
speed of 1500 rpm. The nitro-PAC levels for each set of duplicate samples show good overall
reproducibility, with some exceptions. The effect of engine load on the nitro-PAC extract
concentrations is not repeated for different nitro-PAC. Within the indicated range of error
bars, 2-methyl-1-nitronaphthalene increases in concentration at mid load and thereafter
decreases at higher loads. In contrast, the concentrations of 1-nitronaphthalene, 2-
nitronaphthalene and 1-nitropyrene are more constant across the load range.
Figure 5.10 shows the effect of load on the 1-nitropyrene emissions at the mid-speed of 2500
rpm. The reproducibility is very good and 1-nitropyrene increases up to mid load and then
decreases slightly at high load. Figure 5.11 shows the emissions of nitro-PAC at the highest
speed of 3500 rpm and the three load positions. The nitro-PAC emissions at this high speed
are significantly different with respect to the nitro-PAC emission trends at the lower speeds,
with all nitro-PAC increasing with engine load. The nitronaphthalenes progressively increase
with engine load, whilst 1-nitropyrene is initially stable and then rapidly increases at high load.
Figure 5.12 shows the combined effect of engine speed and load on the emissions of 1-
nitronaphthalene and 1-nitropyrene. The emissions of 1-nitronaphthalene are more stable across
the load range at 1500 rpm, whereas at the higher speed the 1-nitronaphthalene increased with
engine load (Figure 5.12a). Figure 5.12b shows that the greatest variation in the 1-nitropyrene
emissions occurs at high load and for the three speeds.
After establishing the effect of speed and load on the nitro-PAC emissions from the Prima
combustion chamber it is now possible to compare the findings with those from previous
studies, concentrating on 1-nitropyrene. The majority of previous studies have concentrated
on heavy-duty emissions and used dilution tunnel and filter collection.
189
v 1-nitronaphthalene 15.0 <> 2-nitro-1-methylnaphthalene
en * 2-nitronaphthalene w o 1-nitropyrene ..... 12.5 ..... 0 T
E 1o.o y
sz ""-u <> ......
< 7.5
* ~
a_ -I 0 1 1.. 5.0 ...., z ..... 0 2.5 C) [J T
D 0 c .l 0.0
0 20 40 60 80 100
Percentage of Full Load
Figure 5.9 The effect of engine load on the nitro-PAC at 1500 rpm
190
(/) 5 w 1--..... 0 4 C)
E ........
CD 3 c CD
D .... > T a. D 0 2 ..l.. .... ..... c I ..... 1 0 ~
0
C) c 0
0 15 30 45 60 75 90
Percentage of Full Load
Figure 5.10 The effect of engine load on the nitro-PAC at 2500 rpm
191
v 1-nitronaphthalene
40 ~ 2-methyl- 1-nitronaphthalene T en * 2- nitronaphthalene
w o 1-nitr opyrene I I-...... 0 30 en E
........ (J
< 20 0.. I 0 ~ ...., z 10 ...... 0 Cl en c:
0
0 20 40 60 80 100
Percentage of Full Load
Figure 5.1 1 The effect of engine load on the nitro-PAC at 3500 rpm
192
en a) UJ 40 ..... - o 1500 rpm 0 V
01 v 3500 rpm E 30 ....... C) c C)
tU ..&:
20
/ -..&: Q. tU c 0 c ~ - 10 c c c I ..... -0
01 0
c 0 15 30 45 60 75 90
Percentage of Full Load
6 b)
en UJ o 3500 rpm ..... - 5 A 2500 rpm 0
01 E 4 * 1500 rpm
....... C) c t) ~ 3 >-Q. 0 ~ - 2 c I * ..... * - 1:. 0 0 01 c 0
0 15 30 45 60 75 90
Percentage of Full Load
Figure 5.12 The combined effect of speed and load on a) 1-nitronaphthalene and b) 1-nitropyrene emissions.
193
The study by Schuetzle & Perez (1983) on heavy-duty diesels found the 1-nitropyrene extract
concentrations initially increased with load and then severely decreased at high load for a speed
of 2100 rpm. This agrees with the findings of this study at 2500 rpm, with a less severe
decline in the 1-nitropyrene emissions at high load in this study. Schuetzle & Perez also found
the emissions of 1-nitropyrene at mid- and high-loads for a lower speed of 1260 rpm were
more stable in comparison with the emissions at the higher speed, as was the case in this study.
The work also sampled at very low load and 700 rpm, whereupon the highest of 1-nitropyrene
emissions out of all the sampling conditions was generated.
Bechtold et al. (1984) found that the emissions of 1-nitropyrene from a heavy-duty diesel
increased with engine speed whilst under cruising conditions, as did NO. and temperature.
This agrees with the 1-nitropyrene emissions found in this study at mid-load for increasing the
speed from 1500 rpm to 2500 rpm, and also at high load where 1-nitropyrene increases as the
speed was increased.
The study of heavy-duty diesel exhaust wall deposits on road tunnels by Handa et al. (1984)
found that the emissions of 1-nitropyrene were higher at high load compared with low load for
an estimated on-road speed of 1800 rpm. The lack of other engine conditions make it difficult
to gauge the results to those found in this study. By contrast to the road tunnel study the 1-
nitropyrene emissions at 1500 rpm were much more stable in this study, although at higher
speeds the 1-nitropyrene emissions were highest at high load compared with low load.
Draper 1986 compared the nitro-PAC emissions from a heavy duty diesel engine at high load
and moderate speed with the emissions at moderate load and high speed. The authors found
that the nitro-PAC emissions were highest at moderate load and high speed. The fact that the
study relied on only two sampling points combined with both speed and load changing at each
point again makes comparisons difficult. Taking 1500 rpm as a moderate speed and 3500 rpm
as the high speed reveals no correlations between the findings in this study to those by Draper
and eo-workers, except for 1-nitronaphthalene, which was higher at moderate load and high
speed in both studies.
194
Veigl et al. (1994) found that the highest emissions of 1-nitropyrene emitted from a light-duty
diesel were produced over the HWFEf compared with a lower speed and fluctuating load FTP
cycle. The increased nitro-PAC emissions at high speed, would tend to agree with the results
from this study, however transient cycles are difficult to breakdown into whether it is speed
or load which determines the emissions. The authors also profiled 1-nitropyrene emissions over
a range of loads at constant speed of 1560 rpm for a medium-duty diesel. Within the error
bars, some of which were large, the 1-nitropyrene emissions increased up to mid-load
whereafter the emissions decreased slightly, a finding similar to that found at 2500 rpm in this
study but not at 1500 rpm. The authors found the lowest emissions of 1-nitropyrene at low
load and 600 rpm.
The comparison of the findings from this study with previous ones, has in some cases
identified some correlations. However there are also major differences between not only this
study and the other studies, but also amongst the other studies themselves. In most cases, many
more sampling points would be needed to provide a better comparison. The majority of the
studies compared to this study were based on heavy-duty engines. The similarity of heavy-duty
and light-duty combustion in terms of nitro-PAC is unknown. In terms of the range of both
engine conditions and nitro-PAC investigated the results from this study are the most
comprehensive to date for light-duty diesel engines. Another major weakness of the majority
of previous studies is the failing to attempt correlations between the emissions and nitration
parameters, such as NO, and temperature. The next section investigates the importance of such
factors at specific engine conditions.
5.2 I 5 Factors controlling the Nitration in the Combustion Chamber
The close proximity of TESSA to the Prima engine largely removes the post-combustion
reactions simulated by dilution tunnel sampling systems. The short transfer line will allow
some post-combustion reactions to proceed, but not to the same extent as the dilution tunnel.
The greatly reduced post-combustion environment has resulted in much lower levels of nitro
PAC found in this study. This suggests that a significant proportion of nitro-PAC are derived
following combustion processes. Indeed Kittelson et al. (1984) found that nitro-PAC were four
times more likely to be found in the exhaust than in the combustion chamber.
195
The fonnation of nitro-PAC is primarily a function of the concentration of PAH and nitrating
species, tenned NO,. The analysis of a nitro-PAC HPLC fraction by GC/MS NICI identified
that a proportion of PAH present in the fuel had at some stage been converted into nitro-PAC
(Section 5.2.1.2).
To examine the effect of NO, on the nitro-PAC emissions, the NO, concentrations were
examined relative to the nitro-PAC emissions at varying engine conditions (Figure 5.13). Due
to the greater infonnation on nitro-PAC, the speeds corresponding to 1500 rpm and 3500 rpm
are displayed. For all speeds tested, the gaseous NO, concentrations increased with engine load
up to around 80% of full load, and thereafter the NO, levels declined slightly. There is little
correlation between the nitro-PAC emissions with the NO, concentrations at low and mid
engine speeds (R-sq for 1-nitronaphthalene and NO, correlation = 0.327 and p = 0.612, at
low speed.). At the higher engine speed, there is a strong correlation between NO, and the
nitro-PAC (R-sq for 1-nitronaphthalene and NO, correlation = 0.989 and p = 0.067).
The correlation of NO, and nitro-PAC at high speed as opposed to lower speeds, may indicate
that the nitro-PAC formation is temperature controlled to some extent. This suggests that the
nitration reactions are favoured by the higher temperatures existing at high speed. Bechtold et
al. ( 1984) also found a link between exhaust temperatures and the 1-nitropyrene emissions.
It is important to point out that the NO, measurements were derived further down the exhaust
system (Rhead et al. 1991). This work assumes that the overall NO, trends with engine load
are the same close to the combustion chamber as for further down the exhaust. The work of
Pipho et al. (1991) showed the levels of NO, within the combustion chamber were replicated
further down the exhaust pipe.
196
(/)
w 1--0
a)
15.0
12.5
E 1o.o ........ () < 7.5 Q..
I 0 L.. +J
z -5.0
0 2.5 0) c:
(/)
w 1-
-
40
0 30 0')
E ........... () <( 20 Q..
I 0 a-. +J
z 10 -0
0')
c
1250
0 0 :l
1000 ~ :l -.., Ql
750 -0 :l
0
500 -z 0 ><
250 "0 "0 3 i-~----------- T ·------· ..L
0.0 0 20 40 60 80 100
Percentage of Full Load
,. 1-nitronaphthalene • 2-methyl-1-nitronaphthalene
* 2-nitronaphthalene • 1- nitropyrene o NOx
b)
» T/! ,. /
i~~ ..L •
0 20 40 60 80 Percentage of Full Load
100
1000 () 0 :l
800 ° (l)
:l -.., Ql
600 ~-0 :l
0 -400 z 0 ><
200 .:o"0
3
0
Figure 5.13 The investigation of correlations between the nitro-PAC and NOx emissions at a) 1500 rpm and b) 3500 rpm.
197
5.2.1.6 Nitration mechanisms involved
Two nitration mechanisms have been proposed to account for the nitro-PAC emissions from
diesel engines are:
1) Free radical reactions involving NO, and PAH radicals.
2) Electrophilic substitution of PAH, involving N02 in the presence of nitric acid.
(Scheepers & Bos 1992b)
It is important to note that it is only the N02 fraction of NO, involved in the electrophilic
substitution nitration (Lindskog 1983 and Schuetzle 1983), whereas both NO and N02
participate in the free radical nitrations (Scheepers & Bos 1992b). The primary NO, species
formed in the chamber is NO, with N02 existing as a transient species at the flame front,
before being converted to NO in the post-flame region (Arcoumanis 1992). Nitrogen dioxide
can be emitted at low loads due to the conversion of N02 to NO being quenched (Arcoumanis
1992). This translates to N02 contributing ea. 50% to the total NO, emitted at low loads
compared with only 5% at high loads (Campbell et al. 1981, Lenner 1987, Pipho et al. 1991).
The highest concentration of nitric acid, known to greatly enhance nitration, is also generated
at low loads (Harris et al. 1987).
Free radical nitration reactions are favoured by conditions where both the NO, and P AH
radicals are abundant and sufficient potential exist for the reaction to proceed. The high
temperature dependence of free radical reactions would suggest the variables for successful
free radical nitration would exist following combustion and diminish with time thereafter. In
contrast, nitration of PAH by electrophilic substitution with the nitronium ion, N02 +, is not
governed by the same kinetics as free radical processes. Indeed, nitration involving the
nitronium ion (NO+ :J in aqueous solutions will proceed at room temperature (Nielsen 1984 and
Ross et al. 1988). It follows that the greatest potential for electrophilic substitution nitration
occurs at conditions for which nitrogen dioxide and associated nitric acid can interact with
PAH, ie. at the lower exhaust region temperatures.
198
In order to an gain a further insight into the potential of the combustion and initial stages of
exhaust to form nitro-PAC, the nitro-PAC, PAH, and NO, were referred to the amount of fuel
consumed. At low engine speed, l-nitronaphthalene emissions, in terms of the fuel consumed,
are highest at low load (Figure 5.14). The lowest combustion efficiency associated with low
loads result in naphthalene also being highest at low loads (Chapter 4). By contrast, the
gaseous NO, emissions increase with engine load, but the actual contribution of N02 to the
NO, decreases with increased load (assuming a 50% contribution of N02 to NO, at low loads
compared with 5% at high loads). Similar trends are generated for 1-nitropyrene, pyrene, and
NO, emissions at 1500 rpm, with the exception that pyrene increases again at high load (Figure
5.15).
The high emission of nitro-PAC relative to the fuel consumed at low load and low speed, may
indicate that under these engine conditions the greatest free radical interaction with NO. and
PAH occurs (mechanism 1). The other possibility is that low loads result in the nitration of the
surviving PAH by substitution mechanisms involving N02 and nitric acid, via the N02 +
(mechanism 2). The higher N02 and nitric acid combined with the higher survival levels of
PAH (Chapter 4) at low loads, would represent the highest electrophilic substitution nitration
potential during the expansion and exhaust stages (mechanism 2).
Hirakouchi et al. (1990) also found that the highest emissions of l-nitropyrene and PAH,
expressed relative to the fuel burned, were at low loads. The gaseous NO. emissions relative
to the fuel burned peaked at around mid-loads and then stabilised at higher loads, as in this
study. The study using a mini-dilution tunnel found very good correlations with full flow
dilution tunnels.
The dilution tunnel arrangement will favour exhaust electrophilic substitution reactions with
N02 and associated HN03 (mechanism 2). Hence, the similar trends found by Hirakouchi and
eo-workers and by this study provides further evidence that the high level of nitro-PAC at low
loads and low speeds are a consequence of PAH surviving combustion and undergoing
nitration in the early stages of the exhaust/transfer line (mechanism 2). An extended transfer
pipe between the engine and TESSA would enable verification of this theory (Section 6.3).
199
~ a) c ... :::1 0.08 .a
0 0.07
a ....: ' a E Cl c Cl
"ii .s:; -..c Q. ... c 0 !: .2 I
"0 «< c .. :::1 .a
Ci :::1 --0
Cl -"' ' Cl E Cl c ~ ... .s: -.s: Q. ... c
"'0 «< c .. :::1 .Q
"'i :::1 --0
Cl .X: ..... tll .§ )(
0 z
0.06
0.05
0.04
12
11
10
9
8
7
6
40000
30000
20000
10000
0
0
b)
0
cl
15 30 45 60 75 90
Percentage of Full Load
15 30 45 60 75 90
Percentage of Full Load
c NOx
• N02
0 15 30 45 60 75 90
Percentage of Full Load
Figure 5.14 Emissions of 1-nitronaphthalene (a), naphthalene (b), and NOx (c) relative to
the fuel burned at 1500 rpm.
200
Figure 5.15 Emissions of L-nitropyrene (a) , pyrene (b) , and NOx (c) relative to the fuel burned at 1500 rpm.
201
At the higher engine speed of 3500 rpm, the profile of 1-nitronaphthalene and NO, relative to
the total fuel consumed both peak at around 30% of full load (Figure 5.16). This may indicate
that this engine condition represents the optimum conditions for free radical nitration of
naphthalene. The possibility that 1-nitronaphthalene is derived from the electrophilic
substitution of naphthalene by the N02 + is thought unlikely. This is due to the low N02 and
associated nitric acid concentrations that would exist at the temperatures involved with this
load and speed. Secondly, the survival level of naphthalene that would be available is relatively
lower than at the lowest loads.
The 1-nitropyrene emissions at 3500 rpm replicates the proftle of pyrene across the load range
(Figure 5 .17). The close resemblance of the pyrene and 1-nitropyrene emissions relative to the
total fuel consumed at high engine speed, suggests a direct link between pyrene and 1-
nitropyrene at these engine conditions. In fact examination of the percentage recovery of
pyrene relative to the other major PAH at high load and high speed revealed that the pyrene
emissions were probably contributed to by a significant pyrosynthetic input (Section 4.3.2).
The increased mass of pyrene causes a significant input to the formation of 1-nitropyrene,
either by free radical or substitution processes. The possible pyrosynthetic contributions to
pyrene combined with the high temperatures at extreme engine powers, suggest that the
nitration takes place by free radical combustion chamber interactions between NO, and pyrene.
To investigate the nitration potential as a function of engine speed, the profiles of 1-
nitropyrene, pyrene, and NO, relative to the total fuel consumed were plotted at low and high
loads. The emissions of 1-nitropyrene, relative to the fuel consumed, at low load decrease with
speed (Figure 5.18). By contrast, pyrene initially increases with speed and then decreases at
higher speeds. Hence, the greatest interaction between NO,/nitric acid and PAH (mechanism
2) exist at the low temperatures present in the expansion stages of low load. This may indicate
that increased speed at low loads reduces the reaction time available for N02 + and PAH
interactions. At high load, 1-nitropyrene emissions are the reverse to low load, and now
increase with speed (Figure 5.19). As identified, pyrene has a significant input to the nitro
PAC emissions at high speed and high load. The trends at high load suggest that temperature
is an important factor in the free radical formation of nitro-PAC emissions.
202
-.:::1 al " c ...
0.20 ::::11 .a
"'ii 0.18 ::::11 --0 0.16 0
...: ....... 0.14 0, E
" 0.12 c " G 0.10 • -..c a. 0.08 G c 0 ...
0.06 -c 0 15 30 45 60 75 90 I
Percentage of Full Load
bl -.:::1 15.0 " c .... ::::11 .a
"'ii ::::11 -- 10.0 0
0 -"' ....... 0 E 5.0 G' c
" lii • ... • a. IQ 0.0 c
0 15 30 45 60 75 90
Percentage of Full Load
cl 40000
a NOx -.:::1
" c • N02 ... ::::11 .a 30000
" ::::11 --0 20000 0
...: ...... c; E 10000 )(
0 z 0
0 15 30 45 60 75 90
Percentage of Full Load
Figure 5.16 Emissions of 1-nitronaphthalene (a) , naphthalene (b), and NOx (c) relative to the fuel burned at 3500 rpm.
203
-.::1 a)
G 0.012 c. ... ~ .a
0.011 • ~ 0.010 -0
Q 0 .009 .lilt. ...... Q 0.008 _g G c. 0.007 G ,_ >-a. 0 .006 0 ,_ -·c
0.005 I 0 15 30 45 60 75 90
Percentage of Full Load
b)
1.50 -.::1 G c. ,_ 1.25 :J .a
• 1.00 ~ --0 0.75
a .lilt. ...... a 0.50 E ... c. 0.25 G ,_ >-a.
0.00 0 15 30 45 60 75 90
Percentage of Full Load
cl 40000
a NOx -.::1 Cl c. • N02 ,_ ~
.a 30000
• ~ --0 20000 01
..lilt. ...... Q
E 10000 )(
0 z 0
0 15 30 45 60 75 90
Percentage of Full Load
Figure 5.17 Emissions of 1-nitropyrene (a), pyrene (b), and NOx (c) relative to the fuel burned at 3500 rpm.
204
"'0 al • 0.0150 c ... ::I ~
• ::I 0.0125
-0
tit .....: ::::: 0.0100
Q
_g • c • 0.0075 ... >-a. 0 ... -·c:
0.0050 I 1500 2000 2500 3000 3500
Engine Speed (rpm)
1.)
0.300 "'0 • c: ... 0.250 ;;J ~
"'i 0.200 :::J --0 0.150
Q .....: .._
Q
E 0.100
• c 0.050 • ... >-a.
0.000 1500 2000 2500 3000 3500
Engine Speed (rpm)
cl 30000
a NOx ~ • • N02 c 25000 ... :::1 ~
"'i 20000 ::I -- 15000 0
Q ...: -10000 Oi E )( 5000 0 z
0 1500 2000 2500 3000 3500
Engine Speed (rpm)
Figure 5.18 Emissions of 1-nitropyrene (a), pyrene (b), and NOx (c) relative to the fuel burned at low load.
205
Figure 5.19 Emissions of 1-nitropyrene (a), pyrene (b), and NOx (c) relative to the fuel burned at high load.
206
5 2.2 Dinitro-PAC & Njtro-oxy-PAC Profiles for the Prima Diesel Engine
Following the large variations in mononitro-PAC at high speed, the HPLC fractions associated
with the dinitro-PAC elution window were analyzed by GC-ECD for the three loads collected
at high speed (Figure 5.20). There are a great deal of electrophilic compounds present in these
HPLC fractions, and the distribution of the compounds for each load position are similar. This
may indicate that there are common formation routes occurring to produce the compounds
across the load range. At high load the greatest abundance of compounds are evident. This
provides further evidence that high speed and high load favour combustion generation of
secondary nitro-PAC. However, eo-injection techniques could not find any dinitro-PAC. The
dinitro-PAC levels from diesel exhaust are typically much lower than the mononitro-PAC
(Nakagawa et al. 1983, Williams et al. 1986b, and Hayakawa et al. 1992). Given the low
mononitro-PAC levels found in this study, it may not be surprising that the even with the
sensitivity of GC-ECD no dinitro-PAC could be found in the Prima emissions.
The HPLC fractions corresponding to the nitro-oxy-PAC elution window collected at high
speed and for the three loads were also analyzed by GC-ECD (Figure 5.21). The fractions also
contain a vast amount of electrophilic compounds, again with a degree of similarity at each
load, suggesting common origins. No identification of nitro-oxy-PAC were found by eo
injection techniques. The ECD detector was very susceptible to becoming contaminated by the
nitro-oxy-HPLC fractions.
5.3 Oxy-PAC Emissions
The excess air and high temperatures present in diesel combustion enables primary P AC
emissions, such as fluorene, to be converted into oxy-PAC derivatives, in the case of fluorene,
resulting in the formation of primarily fluoren-9-one. It has been proposed that some oxy-PAC
may be responsible for significant mutagenic contributions to diesel extracts (Scheepers & Bos
1992a). In an attempt to gain further information on the potential of diesel combustion to
generate oxy-PAC at different engine conditions, the oxy-PAC HPLC fractions and the polar
fractions from the silica gel clean-up were analyzed by GC/MS in the El mode (Section
3.4.3.2).
207
lOO
~
"' = 0 c. "' ~ ~ u ~
10 lOO
~
"' s::: 0 c. "' ~ ~ ~ u ~
100
100
180
80
180
Temperature ( °C)
(a) Low Load
(b) Mid Load
(c) High Load
300
300
3
Figure 5.20 The gas chromatographic analysis of the dinitro-PAC HPLC fractions collected at 3500 rpm and a) low load, b) mid load, and c) high load by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.2).
208
100
(a) Low
4.1 Load ., = 0 c. ., ~ ~ u ~
10 100 200 300
(b) Mid
4.1 Load ., = 0 c. ., 4.1
~ ~ u r;.il
LOO- 100 200 300
(c) High
4.1 Load ., = 0 c. ., 4.1 ~ ~ u r;.il
100 200 300
Temperature ( °C)
Figure 5.21 The analysis of the nitro-oxy-PAC HPLC fractions collected at 3500 rpm and a) low load, b) mid load, and c) high load by gas chromatography with electron capture detection (Chromatographic conditions given in Section 3.4.3.1.4.3).
209
In order to obtain a preliminary appraisal of the effect of engine conditions, the oxy-PAC
HPLC fractions corresponding to 1500 rpm and low load were compared with the same
fractions obtained at 3500 rpm and high load (Figure 5.22). The analysis shows few major
peaks between the retention index markers. Presently, no identifications of the larger peaks
have been ascertained. However, there is little difference between the two extremes of engine
conditions. The preliminary profiles suggest that the pathways which lead to oxy-PAC and
polar PAC formation are at least similar across many engine conditions. The exact structures
and levels of compounds under different engine conditions requires further work, possibly
using LC/MS (Galceran & Mayona 1994) or LC/GC (Kelly et al. 1992); techniques more
suited to the more polar oxy-PAC.
The comparison of the polar methanol fractions from the silica gel clean-up at low and high
engine power also shows a series of small peaks, with a few prominent ones interspaced
between the retention index markers (Figure 5.23). There appears to be some matching of
peaks between the two samples, suggesting that the compounds are derived from similar
pathways. As for the oxy-PAC HPLC fractions, no firm identifications have yet to be made,
and further analysis by LC/MS or LC/GC would be useful.
210
Key Retention Index markers: A = Naphthalene B = Phenanthrene
lOO .
~ Col
= = "0
= = .Q 5 < ~ .:: .... = "ii ~
lOO
~ Col
= = "0
= = 50 .Q
< ~ .:: -= "ii ~
A
3.40
3.40
C = Chrysene D = Benzo(ghi)perylene
11.06 18.32 25.58
Total Ion Chromatogram (TIC):
c
(a) Low
Power
D
33.24 40.50 48.16
Retention Time (minutes) ·
B Total Ion Chromatogram (TIC):
(b) High
Power
D
11.06 18.32 25.58 33.24 40.50 48.16
Retention Time (minutes)
Figure 5.22 The analysis of the oxy-PAC HPLC fractions ofTES collected at low engine power (a) and high engine power (b) by gas chromatography/mass spectrometry operated in the electron impact mode (Chromatographic conditions given in Section 3.4.3.2) .
211
Key
Retention Index markers: A = Naphthalene B = Phenanthrene
100 A
I .Ll ~~
j
C = Chrsyene 0 = Benzo(ghi)perylene
Total Ion Chromatogram (TIC): B
c
J J.l l - l
J>
(a) L ow ower p
-3.41 11.09 18.37 26.05 33.33 41.01 48.29
lOO
A
3.41
Retention Time (minutes)
B Total Ion Chromatogram (TIC):
c
I
D
(b) High
Power
-'---'-'-------11.08 18.35 26.02 33.28 40.54 48.20
Retention Time (minutes)
Figure 5.23 The analysis of the polar fractions from the silica gel clean-up of TES collected at low engine power (a) and high engine power (b) by gas chromatography/mass spectrometry operated in the electron impact mode (Chromatographic conditions given in Section 3.4.3.2).
212
54 Summary of Findings for Secondary Nitro- and Oxy-PAC Emissions
1) Transformation of primary organic emissions to secondary emissions occurs in the
combustion chamber. Mononitro-PAC were present in TES but not in the fuel nor oil.
2) Mononitro-PAC were found at low levels for all exhaust samples, analysed by GC-ECD and
GC/MS NICI. The levels were much lower than those found with dilution tunnel sampling
systems and as reported in the literature. This may mean that dilution tunnel/filter sampling
results in enhanced nitration as a consequence of greater air dilution. Alternatively, the dilution
tunnel sampling technique does not replicate the tailpipe and environment processes and instead
leads to high nitro-PAC by artefact processes.
3) The GC/MS NICI results indicate that a proportion of the fuel after combustion undergoes
nitration resulting in the exhaust distribution of the nitro-PAC derivatives similar to the PAH
precursors in the fuel.
4) Different proflles of nitro-PAC with respect to engine load are found under varying speed
regimes. Hence, it is not sufficient to rely on specific engine sampling positions to establish
the processes controlling nitration reactions. In this study a full range of speeds and loads has
for the first time been achieved for a light-duty diesel engine.
5) The concentration of mononitro-PAC at low- and mid-speeds did not correlate with gaseous
NO. concentrations. High speed resulted in an increase in the mononitro-PAC and NO.
concentrations with engine load, possibly indicating that the nitration is dependant on high
temperature reactions, such as those involved with free radical reactions.
6) In terms per gram of fuel burned the highest nitro-PAC emissions were at low load under
low engine speed, as were the PAH emissions. Hence, the conditions at low loads and low
speed favour low temperature interactions of N02, in the presence of nitric acid, via N02 +,
and P AH that have survived to the exhaust.
213
7) At high load and high speed, the formation of l-nitropyrene is enhanced by the pyrene mass
being increased, possibly through pyrosynthetic processes. The peak of 1-nitronaphthalene and
NO. at around 30% of full load suggest that free radical processes may be enhanced for l
nitronaphthalene at these engine conditions. Hence, the high temperature environment at high
speed and high load encourages free radical nitration reactions.
8) The analysis of the dinitro- and nitro-oxy-HPLC fractions by GC-ECD could not find any
evidence for the presence of such compounds. The profiles at low-, mid-, and high-loads for
the top speed of 3500 rpm were similar for both sets of dinitro- and nitro-oxy-PAC HPLC
fractions. This shows that similar reactions may be occurring across the load range at high
speed. Further analysis by GC/MS NICI may enable identifications of other such structures.
9) The preliminary search for oxy-PAC failed to make any positive identifications of
structures. The profiles were similar at low and high engine power settings, suggesting that
the pathways to polar PAC may be similar for both power settings. Further analysis is
required, possibly using LC/MS.
214
Chapter 6-Final Discussions, Conclusions, & Future Work
This chapter is divided into discussing the major findings from both the primary and secondary
emission chapters (Section 6.1), listing the overall conclusions of the work (Section 6.2), and
finally suggesting future work (Section 6.3).
6.1 Final Discussions
The main purpose of this research was to produce emissions maps of specific classes of organic
compounds for a range of speeds and loads. Emissions maps provide useful information for
engineering and fuel companies alike. Such information will become increasingly important
if legislation requires specific chemical criteria to be established over a range of speed and load
test conditions before awarding engine certification.
The emission profiling was performed on a relatively modern naturally aspirated Perkins Prima
DJ light-duty diesel engine using the TESSA to sample the emissions close to the combustion
chamber. The Prima was sampled for a total of 26 different speeds and loads and the samples
characterised for the primary organic emissions and secondary emissions, using two versions
of TESSA (fable 6.1).
T bl 6 1 0 a e vervtew o t e miSSIOn ro ll! fhE' ' PfilinR esearc hP rogramme
Speed Percentage of Full Load TESSA Primary Emissions Secondary (rpm) version profiled Emissions profiled
1000 1, 25, 50, 75, & 100 Initial n-alkanes, PAH, PASH, -& Alkyl-PAC
LOOO 1,8, 17, &25 Upgraded Fluorene -
1500 1, 50, & 85 Upgraded n-alkanes, PAH, PASH, Nitro-PAC & Oxy-Alkyi-PAC PAC
2000 I , 25, 50, 75, & 100 Initial n-alkanes, PAH, PASH , -& Alkyl-PAC
2500 I , 50, & 85 Upgraded PAH, PASH, Alkyl-PAC 1-Nitropyrene
3000 I , 25, 50, 75, & 100 Initial n-alkanes, PAH, PASH , -&Alkyl-PAC
3500 I , 29, &90 Upgraded n-alkanes , PAH, PASH , Nitro-PAC, Oxy-_AJ!yl-PAC PAC. & polars
215
Sampling with both the initial and the upgraded TESSA versions was found to be both rapid
and reproducible (Sections 3.2.1, 3.3, and 3.4.1). The profiling of the primary organics was
successfully achieved with a simple work-up and well established detection systems (Section
3.2). The secondary emissions were focused on the nitro-PAC, for which more complex work
up and specialised detection system were required (Section 3.4). The identification of nitro
PAC by GC-ECD and GC/MS NICI provided detailed quantitative measurements. The
technique of reverse phase HPLC with fluorescence/chemiluminescence detection, with on-line
pre-column reduction, was proved a valuable future detection technique.
The findings from both the primary PAH and n-alkane emissions and the secondary nitro- and
oxy-PAC emissions were presented and discussed in Chapters 4 & 5 respectively. The
following sections discuss further the major findings from those two chapters and investigate
both the wider environmental implications and appropriate control strategies implied from the
results. The discussions are divided into those relating to the primary organics (Section 6.1. l)
and the secondary organic (Section 6.1.2) emissions.
6.1.1 Primary Organic Emissions
One of the major findings from the primary organic emission profiling was the close
resemblance of the aliphatic and aromatic fractions of emissions to those of the same fractions
in the diesel fuel (Section 4.1). This provides strong evidence for fuel survival being a major
contributory factor in the emissions composition (Andrews et al. 1983, Williams et al. 1986
& 1989, Barbella et al. 1988 & 1989, and Zeijewski et al. 1991). The highest levels of
survival were associated with low loads, in agreement with previous light-duty DI studies
(Andrews et all983, Williams et al. 1986, Barbella et al. 1989, and Zeijewski et al. 1991)
(Section 4.2.2.). In this study, the highest survivals of all were obtained at low load and low
speed. This highlights the engine conditions at which light-duty DI engines are susceptible to
poor combustion. Similar conclusions have been reached by a number of other researchers
(Mills et a/.1984, Yoshida et al. 1986, and Assaumi 1992). The reasons for the inefficient
combustion at low loads for the DI engine, comes from the low temperatures and insufficient
air:fuel interactions. Further evidence that temperature on its own, has a strong effect on the
combustion efficiency was illustrated by sampling the Prima with and without engine
216
conditioning (conditioning normally consists of 1 hour at 3000 rpm and 90 Nm followed by
a further fifteen minute conditioning at the sampling position) (Section 4.3.1). The
unconditioned sampling resulted in a higher survival of fluorene (1.31%) compared with
conditioned sample (0.96%). It follows, that the higher emissions associated with the cold start
were a result of the lower combustion chamber temperatures and greater combustion instability
(Asou et al. 1992). The Earth Resources Research estimated that cold start conditions
contributed approximately one third of the total volatile organic compound emissions from
passenger cars in 1990 (Holman et al. 1993 cited by QUARG 1993).
The predicted rise in passenger diesels in the urban environment (QUARG 1993 and
Parliamentary Office of Science & Technology 1994) and the increasing reliance on the DI,
rather than IDI fuel injection configurations (Bulmer 1990, Diirnholz et al. 1992, and
Belardini et al. 1993) may be of significance to the urban environments. The inefficient DI
diesel combustion occurs at the low speed and low load stop-start driving associated with
congested city centres. The average car journey times in the UK are low with about 40% less
than three miles (QUARG 1993). The short journey time would produce relatively higher
emissions due to the low temperature existing during such driving periods (Martin and Shock
1989 cited by QUARG 1993). The cold-starting problem is more severe for petrol engines
(QUARG 1993), but it follows that for very short trips, emissions from diesel engines would
by aggravated by reduced ambient temperatures (Tritthart et al. 1993).
The similar recoveries experienced for different PAC and n-alkanes at each load position
sampled at 1000 rpm, suggested that, for this low speed, the fuel survives combustion largely
unchanged (Section 4.2.2). The n-alkanes had lower survivals, compared with those PAC with
corresponding boiling points, suggesting that the straight chain-n-alkane structure is relatively
easier to break down than that of the more stable ring PAC structures. The PAC may be
supplemented by a greater combustion generation input (Section 4.2.5). Combustion studies
using simple aliphatic fuels have been found to result in PAC emissions (Crittenden & Long
1973 and Cole et al. 1984). The higher PAC emissions found in this study may be a result of
the higher combustibility of the aliphatics, especially in low temperature and oxygen zones
needed for complete combustion of aromatics, and also by combustion generation of PAC
217
involving aliphatic radicals. The extent to which preferential survival and/or combustion
chamber reactions contribute to the emissions at different engine conditions might be revealed
in further 14C labelling experiments on the Prima at the University of Plymouth.
The similarity of the emission and fuel proftles means that the fuel composition will have an
important role in determining the emissions entering the environment. The primary organic
emissions, sampled close to the combustion chamber, found in this study replicate the trends
with respect to speed and load as have been found by previous studies. However, there may
be changes in the chemical composition of the exhaust further down the exhaust pipe not
included in this study. Hayona et al. (1985) found that the transfer of combustion formed PAH
into the exhaust caused a large decrease in the PAH emitted. By fitting an extended 3 m length
of transfer pipe between an IDI Ricardo Hydra and TESSA, Trier (1988) found that the
aliphatics were preferentially burned-out compared with the aromatic and polar constituents
of fuel surviving combustion. Hence, of the fuel surviving combustion it is the more hazardous
P AC which enter the environment compared with the much less harmful n-alkanes.
There are many thousands of PAC present in diesel fuel and only a selection of these have
been evaluated for their mutagenic and carcinogenic effects (IARC 1989). Of those tested, the
methyl-derivatives of fluorene, phenanthrene, anthracene, pyrene, chrysene, and
methylbenzo(a)pyrene have been found to be carcinogenic (Melikian er al. 1983, IARC 1989,
and Scheepers & Bos 1992b). The concentrations of the methyl derivatives of two to four ring
PAH are higher in diesel fuels than such carcinogenic PAH as benzo(a)pyrene, and it may be
that alkyl-PAR play a important part in the hazard associated with diesel emissions (Scheepers
& Bos 1992b). Within each group of methyl derivatives, the precise position of the methyl
group on the PAC ring structure is key to the potency of the compound. For example, the bay
regions associated with 5-methylchrysene result in a higher carcinogenic potency than 6-
methylchrysene (Melikian et al. 1983 and IARC 1989).
By contrast to the poor combustion at low loads, the combustion efficiency was found to be
greatest under high loads (Section 4.2.4 & 4.3.2). Similar conclusion have been reached by
Andrews et al. (1983), Williams er al. (1986), Barbella et al. (1988), and Zeijewski et al.
218
(1991). Thus the DI system works effectively under high fuelling conditions, such as those
encountered with steep gradients. However, in terms of the environment, high loads will result
in a higher output of possible mutagens present in the fuel, such as benz(a)pyrene and the
mutagenic alkyl-PAH, purely as a consequence of the greater fuel consumption.
The highest combustion efficiencies for organic emissions were associated with mid-speeds,
between 2000 rpm and 2500 rpm, for low loads (Section 4.3.3). At higher loads, the effect
of speed was minimal. Zeijewski et al. (1991) also correlated mid-speeds with the optimum
swirl, reaction time, and temperatures for efficient combustion/power to develop. As with high
loads, the greater fuel consumption at high speeds would result in the highest output of
organics. Thus for high load sections of high-speed motoring, the greatest emissions of all will
be produced, but the scope for improvement, in terms of combustion efficiency, is not as great
as at low loads.
The large proportion of organic emissions derived from fuel surviving the combustion process
and the extent to which different engine conditions affect the extent of the survival is of great
significance to control strategies. Firstly, the lowest combustion efficiencies associated with
low loads highlight the shortcomings of the Prima fuel injection system. In order to enhance
combustion at low loads, a greater air:fuel interaction is required combined with higher
temperatures (Dent 1980, Kamimoto et al. 1980, Karimi 1989, Walsh & Bradow 1991, &
Konno et al. 1992b & 1993). Higher fuel injection pressures, increased fuel injector outlets,
valve covered orifice injectors, reduced crevice volumes and improved combustion chamber
geometry (including more radical designs to encourage turbulence) have all been shown to
lower UHC emissions (Ning et al. 1991, Walsh & Bradow 1991, Akaska & Tamanouchi 1992,
Kobayashi et al. 1992, Konno et al. 1992b & 1993, and Poulton 1994). It is vital that an
integrated approach is adopted, to prevent specific improvements in some emissions at the
expense of others. Increasingly, computer technology is being used to control the different
systems and control the combustion process, allowing the optimum conditions for the lowest
emissions to be achieved at varying engine conditions (Carroll 1991, Walsh & Bradow 1991,
Amstutz & Del Re 1992, Bazari & French 1993, Ladommatos et al. 1993).
219
The close resemblance of the UHC emissions and primary organic emissions (Sections 4.2.2
& 4.2.3), may enable a significant reduction of the organic emissions to be achieved through
improvements to the Prima engine. With respect to the Prima engine design, recent
improvements have included two-stage injectors and turbocharging.
The two-stage injectors were introduced primarily to reduce noise, but may also reduce UHCs
as a consequence of less fuel impingement on the chamber walls and improved spray
characteristics (Shakal & Martin 1990, Arcoumanis et al. 1993, and Konno et al. 1993).
Abbass et al. (1991b) studied the effect of turbocharged and natural aspiration on the organic
emissions from a 4 I DI diesel. The authors found similar profiles for both engines, but with
higher recoveries of exhaust PAH for the turbo version. This may reflect the higher
temperatures and pressures associated with turbocharging, leading to relatively more
pyrosynthesis compared with the naturally aspirated version. The authors point out that, sump
oil and exhaust/engine deposits may have been responsible. Following the findings from
Abbass et al. (1991b) the turbo version of Prima, may reduce overall UHCs whilst promoting
specific combustion reactions under certain engine conditions. The effect of the improvements
on the combustion of organics could be achieved by comparing the profiles of the up-to-date
Prima using TESSA, with those presented in this study (Section 6.3).
The other control strategy that might reduce emissions, derived predominately from fuel
survival, is manipulation of the fuel composition. By altering the fuel composition, for
example, removing identified mutagenic components, it follows that the emissions would be
improved. A great deal of research has attempted to identify the properties of fuels which
determine the pollutants (Weaver et al. 1986, Yoshida et al. 1986, Miyamoto et al. 1992, Roj
1992, Naber et al. 1993, Ryan & Erwin, 1993, and Stradling et al. 1993). One of the major
findings from these studies, was the link between particulate emissions and the aromatic
content of fuels. Some studies however, dispute this, and propose that the link is a
consequence of fuel density (Weidmann et al. 1988, Cowley et al. 1993, FIHysand et al. 1993,
and Tritthart et al. 1993). The possible link between particulates and fuel aromatics led to the
US state of California imposing a 10% upper limit on the aromatic content of diesel fuel,
whilst in Sweden there are three specifications: a limit of 5% in class I, 20% in class II, and
220
25% in class Ill (Cooper et al. 1993 and Nikanjam 1993). Such limits require substantial
changes within the petrochemical industry to enable the hydrotreatment processes, capable of
reducing the sulphur and aromatics, to be employed (Weaver et al. 1986 and Martin & Bigeard
1992). However, the introduction of new fuel blends enables the benefits to be realised
immediately. Improvements in engineering designs take time before the new designs filter
down into the car market. Severe changes in fuel composition can cause problems in terms of
fuel pumps and associated hardware (Roj 1992 and Noble 1994).
The increased range of percentage recoveries for the different primary organics, such as
fluorene compared with pyrene, at specific engine conditions, suggest that fuel survival
unchanged is not solely responsible for the emissions. Thus at high speed and low load, it may
be that at a greater range of temperatures as well as air: fuel ratios are generated in the chamber
(Section 4.2.3). This may lead to some organics surviving relative to others, or alternatively
the great range of temperatures may promote combustion chamber reaction pathways. Whether
it is preferential survival and/or combustion generation which are favoured at specific engine
conditions; improved air:fuel mixing and higher temperatures at low loads may reduce the
emissions. A higher overall combustion chamber temperature confined within a narrow
temperature gradient would increase the chance that different structures would be combusted
to the same extent. Advanced injection would also provide higher temperatures and a greater
reaction time, thus reducing UHCs and particulates, but possibly at the expense of higher NO.
emissions (Zelenka et al. 1990 and Zhang et al. 1993). For this reason, a computer controlled
injection system, which could control not only the precise quantity of fuel injected but also the
time at which the injection started, is essential. Such fuel injectors are constantly under
development and recently Cummins Ltd. have developed a fuel injector with infinitely variable
fuel metering and injection timing for heavy-duty diesel engines (Bunting 1992).
The influence of methyl-derivatives on the environment may be reduced at high speeds and low
loads, since under such conditions emissions of the methyl-PAC are lowered relative to the
unsubstituted PAC (Section 4.2.6). The overall hazard of the emissions may be reduced at
specific engine conditions as a consequence of the reduction in the methyl derivatives, some
of which are more hazardous than the non-substituted PAC. The possibility that pyrosynthetic
221
reactions may be active under such engine conditions may result in the formation of more
hazardous comJxmnds. Fuel rich zones combined with the range of temperatures that may exist
at low load and high speed may present the optimum conditions for pyrolysis reactions
(Burrows & Lindsey 1961, Badger et al. 1964, and Tancel1 et al. 1995b). The work of both
Burrows & Lindsey (1961) and Badger et al. (1964) showed that the greatest pyrolysis
products from simple fuels were obtained at a optimum mid-range temperatures.
For instance, Burrows & Lindsey (1961) found that the greatest quantities of naphthalene,
acenaphthylene, phenanthrene, pyrene, fluoranthene, anthracene, and coronene were formed
at ea. 850°C (with temperatures tested ranging from 700°C to 1000°C} for ethylene and ea.
1070°C (with temperatures tested ranging from 900°C to l100°C} for methane flames. The
increase in the relatively lower molecular n-alkanes relative to the higher weight n-alkanes may
be evidence for cracking reactions being promoted under such conditions (Nelson 1989). Thus
the conditions of high speed and low loads may represent a wide range of temperatures, which
allow specific combustion generations to occur to a greater extent than may be possible with
the narrow range of temperatures associated with other engine conditions.
The extremes of 3500 rpm and close to full load resulted in a much higher recovery of pyrene
relative to the other organics; this provided strong evidence that the high temperature
conditions associated with these engine conditions resulted in significant combustion generation
of pyrene. At other engine conditions, the yield of pyrene appeared atypical compared with
other PAC. The 14C-pyrene research by Tancell (1995a) has shown that the structure of pyrene
is more easily pyrosynthesised than smaller ring PAC. Pyrene may be of significance to the
formation of secondary organics emissions, specifically 1-nitropyrene, in the combustion
chamber (Section 6.1.2).
The measurement of primary emissions shows that there is a common trend of fuel survival
occurring to different extents depending on the temperature, reaction time, and air:fuel
interactions associated with different engine conditions. Combustion generation reactions at
specific engine conditions may also influence the emissions.
222
The situation can be shown as in Figure 6.1. By dividing the combustion chamber into
different zones of temperature and time the effect of different speeds and loads can be
envisaged. Thus at high loads the range of temperatures is narrow and hence compounds either
survive in the low temperature zone (Z..) or are completely combusted in the high temperature
zone (ZJ (Figure 6.1a). At low loads, a greater range of temperatures may exist in different
temperature zones of the chamber, enabling differing combustion efficiencies and combustion
reactions to proceed at specific temperatures (Figure 6.1b). The time available for combustion
reactions will also effect the emissions, such that higher speeds will reduce the time available
for total combustion to occur, thus allowing some compounds to undergo pyrosynthetic
reactions rather than complete combustion.
6 I 2 Secondary Organic Emissions
The proflling of some of the highly mutagenic and possibly carcinogenic nitro-PAC was the
ultimate aim of this study. The results relating to the mononitro-PAC, presented in Chapter
5 are unique in two ways. Firstly, the study covers a greater range of speeds and loads than
published to date. Secondly, and more importantly, the study utilises the unique TESSA
developed at the University of Plymouth. The close proximity of TESSA to the Prima engine
allows the contribution of combustion generation to nitro-PAC emissions to be examined over
different engine running conditions.
One of most important findings from the oitro-PAC proflling on the Prima using TESSA, was
that the levels of mononitro-PAC emitted (average extract concentration of 1-nitropyrene of
5.3 ppm) were much lower compared with previous studies which relied on dilution
tunnel/filter sampling systems (average of 1-nitropyrene in extracts of 627.9 ppm) (Section
5.2.1.3). The lower monooitro-PAC levels shortly after combustion agree with the levels (4.5
ppm of 1-nitropyrene in extract) found by sampling diesel exhaust with a cyclone particulate
separator (Li & Westerholm 1994). This implies that extended exposure of primary organics
to dilution enhances the formation of nitro-PAC. The increased nitro-PAC emissions found
from dilution tunnels may in some cases be a consequence of the artefact potential associated
with the filter collection used in such systems. In contrast, the TESSA design minimises the
opportunity for artefact formation.
223
High Load
Key
Z c = Complete combustion zone
Zi = Intermediate combustion zone
Zu = Unburnt fuel zone
Low Load & High Speed
Figure 6. 1 Concept of different temperature and time zones necessary for complete combustion. High loads result in the highest temperatures , resulting in a large complete combustion zone (ZJ, and a much smaller zone in which fuel survives intact (ZJ. The gradient in terms of the temperature and time existing in the intermediate combustion zone (Zj is narrow. At low load and high speed the unburnt fuel zone is greater and the range of temperatures and time available for combustion is large, enabling preferential survival and/or combustion chamber generation of PAC.
224
The work-up procedure was also found not to produce artefacts. The nitro-PAC found in this
study confirm that such compounds can be formed during combustion. Kittleson and eo
researchers (1984) found that 1-nitropyrene was formed in the combustion chamber, but not
to the same extent as in the exhaust.
Studies using dilution tunnel sampling have estimated that nitro-PAC can make significant
contributions to the direct-acting mutagenicity associated with diesel extracts. For example,
Schuetlze (1983) quoted that 66% of the total direct-acting mutagenicity was attributable to
seven nitro-PAC. Other studies have put the contribution of 1-nitropyrene on its own, at
between 10 to 40% of the total mutagenicity of diesel extracts (Schuetzle et al. 1981,
Nak:agawa et al. 1983, Schuetzle & Frazier 1986, & Veigl et al. 1994). The levels of 1-
nitropyrene found in this study would put the mutagenic contributions significantly lower and
more in line with on-road measurements of 1-nitropyrene (Gorse et al. 1983). Thus a key
question in both terms of the nitro-PAC emissions and the associated mutagenicity centres on
whether post-combustion reactions are responsible for nitro-PAC emissions.
The regular profile of nitro-PAC emissions at different speeds and loads suggesting that the
conversion of fuel components occurs across a range of engine conditions (Sections 5.1 &
5.1.1) may be significant to controlling nitro-PAC emissions. The similar distributions of the
nitronaphthalenes in the emissions compared with the distribution of the naphthalenes in the
fuel provides further evidence for the fuel influencing the nitro-PAC emissions (Section
5.2.1.2).
The trends of nitro-PAC with respect to speeds and loads found in this study (Section 5.2.1.4)
are comparable with previous findings. For example, at a mid-speed of 2500 rpm the increase
in 1-nitropyrene emissions with engine load, followed by a decline at higher loads was also
found by Schuetzle & Perez (1983) at 2100 rpm, and by Veigl et al. (1994) at 1560 rpm. The
two studies also found 1-nitropyrene emissions to be more stable across the load range at lower
speeds. Bechtold et al. (1984) found 1-nitropyrene to increase with speed, as did the 1-
nitropyrene emissions at mid-loads and high loads in this study. Handa et al. (1984) found that
the 1-nitropyrene emissions in road tunnels were highest at high loads compared with low
225
loads for a speed of 1800 rpm. A much higher speed of 3500 rpm, produced the highest nitro
PAC at high loads, in this study. There was a correlation between the higher levels of 1-
nitronaphthalene at moderate load and high speed compared with high low and moderate speed
in this study and that undertaken by Draper (1986).
The correlations are complicated by the studies being based on medium- and heavy-duty
engines, whereas this study is based on a light-duty diesel. This may explain why Veigl and
eo-researchers (1994) found the same 1-nitropyrene emission trends to this study at a lower
speed of 1560 rpm, and similarly the road tunnel study by Handa et al. (1984) found the same
effects as found in this study at high speed, but at a much lower speed of 1800 rpm. However,
there remain large discrepancies between some studies. Schuetzle & Perez (1983) found the
highest levels of 1-nitropyrene at low load and 700 rpm, whereas Veigl et al. (1994) found the
lowest emissions at low load and 600 rpm. This may reflect greater artefact contributions at
the low temperatures associated with low engine power in the early study, which may not have
occurred in the later study (even though no direct evaluation of artefact formation is reported
by Veigl et al. 1994).
The range of steady state conditions sampled in this study for the light-duty Prima engine are
the most comprehensive to date, and allow a greater insight into the nitration process under
different driving conditions. The results show that high speed and high load motoring would
result in the highest contribution of nitro-PAC to diesel extracts, whereas lower speeds more
representative of urban areas, would be relatively enriched by PAH surviving combustion. In
terms, of the nitro-PAC emissions emitted per gram of fuel burned, the greatest nitro-PAC
would be emitted at low loads for low speeds. At high speed the greatest nitro-PAC emissions
would be at higher loads (Section 5.2.1.6).
The correlation between the nitro-PAC, especially 1-nitronaphthalene, and the gaseous NO,
emissions at 3500 rpm shows that specific engine conditions favoured the nitration reactions
(Section 5.2.1.5). At the high speed of 3500 rpm, the associated temperatures may provide
sufficient temperatures for the free radical combustion formation of nitro-PAC to proceed. In
the case of 1-nitropyrene, a much greater supply of pyrene to the nitration process was
226
generated at high speed and high load, possibly also as a result of increased pyrosynthetic
contributions (Section 4.3.2). The highest emission of 1-nitropyrene at high speed and high
load, correlated with the greater pyrene available for nitration. At high engine powers the
formation of 1-nitropyrene is controlled by the mass of available pyrene.
Thus the parameters controlling the nitration are the NO., PAH, and temperatures available
at specific engine conditions. The correlation with NOx and also temperature was also made
for 1-nitropyrene emissions from a heavy-duty diesel (Bechtold et al. 1984). The formation
of nitro-PAC by post-combustion reaction of N02 in the presence of nitric acid, involving the
nitronium ion, would be favoured at the lower exhaust temperatures associated with low loads
and speeds. This arises from both the higher levels of N02 and nitric acid in the exhaust, under
low loads and speeds, and the greater primary PAH emissions which also survive combustion
(Section 5.2.1.6).
The results from this study suggest two control strategies; firstly by the reduction of the PAH
content of the fuel, such as the naphthalenes and pyrene by hydrotreatment, would remove the
fuel survival source for nitration at low loads and speeds. There has been relatively little
research into establishing whether different fuel blends can reduce nitro-PAC emissions. Saito
et al. (1982) found higher levels of 1-nitropyrene per mile travelled with high cetane index,
low specific gravity, and high top end distillations point fuels. The work of Saito et al. also
showed that improved fuel injection systems resulted in both reduced overall UHC's and also
1-nitropyrene. Henderson et al. (1982, 1983, & 1984) found that a proportion of the fuel
underwent nitration, whereas Shore (1986) could fmd no link between different fuel blends and
the emission of 1-nitropyrene. Shore did show that DI engines emit significantly more 1-
nitropyrene than the IDI version of the same engine, for most of the fuels tested. For example,
testing of the standard US diesel produced 11.1 Jtg/hour of 1-nitropyrene for the DI compared
with only 0. 78 Jtg/hour for the ID I. As the author notes this was not attributable to NO., since
the DI typically produced three to four times more NOx than did the IDI.
The increased pyrosynthetic activity associated with the formation of both pyrene and 1-
nitropyrene at high speeds may result in fuel measures being ineffective, and under these
conditions a second control strategy, employing post-combustion control measures may be
227
necessary. Post-combustion measures, such as using oxidative catalysts, significantly reduce
PAH precursors to nitro-PAH (Andrews et al. 1987, Henk et al. 1992, Davies et al. 1993,
Floysand et al. 1993, Zelenka et al. 1990, and Zelenka & Herzog 1993).
Research by Veigl et al. (1994) showed that oxidative catalysts lowered the emissions of 1-
nitropyrene from a light-duty engine over the HWFET and FTP transient test cycles. The
authors also found that the addition of fuel additives reduced the 1-nitropyrene emissions. In
contrast, the study by Draper et al. (1988) found that the addition of a barium fuel additive
aimed at suppressing smoke formation, resulted in significant increases of PAC emissions.
Scheepers et al. (1994) found low levels of 1-nitropyrene emissions from light-duty diesels
with oxidation catalysts, however how much of this was a result of the catalyst not possible
to determine, since no comparative studies on the same engine without catalysts were
performed.
The hardest nitration parameter to reduce is that of NO, emissions. This is due to NO,
emissions being related to the peak temperature outputs, and hence the combustion efficiency
itself (Fenn 1982). However, retardation of fuel injection using computer controlled systems
can lower NO, levels without causing other pollutants, typically particulates to rise. Schuetzle
& Perez (1986) showed that retarding fuel injection reduced not only NO, but also the level
of 1-nitropyrene emitted. Pilot injection has also been found to lower NO, emissions (Shakal
& Martin 1990). Another method used to reduce NO, is to re-circulate and mix the exhaust
gases with the fresh inlet air going into the combustion chamber, a technique known as exhaust
gas recirculation (EGR). The NO, is reduced as a consequence of lower temperatures and
reduced oxygen levels (Amstutz & Del1992 and Diirnholz et al. 1992). As with fuel injection
retardation, an integrated approach would be needed so as not to increase particulates and
UHC's. Veigl et al. (1994) found that EGR reduced the levels of 1-nitropyrene over the high
temperature HWFET cycle but reduced much less the 1-nitropyrene emissions (and in one case
the EGR actually resulted in higher 1-nitropyrene emissions) over the more varied and
generally lower speed FTP cycle. This supports the idea that there are two modes of nitration,
one favoured at high temperatures and the other at lower temperatures. Hence, EGR reduced
the overall combustion temperatures, and in so doing increased the lower temperature N02 &
associated HN03 nitration mechanism, whilst limiting combustion chamber nitrations.
228
Specially designed combustion chambers such as the combustion chamber for disturbance
(CCD) described by Konno et al. (1993) can be used to simultaneously reduce NO. and
particulates by two stage combustion; in a much more controlled way than is feasible with fuel
injection retardation. The NO. levels may also be reduced by reductive catalysts, commonly
using copper ion-exchanged zeolites, which can reduce NO, even in the presence of oxygen
(Konno et al. 1992a).
The realisation that reduction catalysts in diesel engines require enhancing the hydrocarbon
input to the exhaust (thus requiring an additional oxidative catalyst) for the reaction to proceed,
and the narrow reduction operating temperatures, has led to investigating selective non
catalytic reduction (SNCR) or thermal deNO, techniques (Andrews & Gibbs 1993). The
process uses chemical injections of gaseous HNCO or ammonia to react with NO (the major
constituent of NOJ to form nitrogen, water, and in some cases carbon dioxide (Andrews &
Gibbs 1993). The extent to which NO, reductions achieved by employing novel combustion
chamber designs, reductive catalysts, and SNCR effect the levels of nitro-PAC remains
uncharted work.
It is important to point out that diesel engines are not the only source of nitro-PAC emissions
in the atmosphere. Nitro-PAC can also be formed by atmospheric reactions with primary
organics, such as pyrene (Pitts 1985, Arey 1986 &1987, Atkinson 1987a,b, & c, 1990a&b,
Zeilinska et al. 1989a&b, Helmig et al. 1992, Ciccioli et al. 1993). These reactions form
isomers of nitro-PAC not produced by diesels, for example 2-nitropyrene. However, since
diesel engines in the urban environment will be in close proximity to human biological systems
the direct emissions of nitro-PAC will be important over the limited time that such compounds
can exist in the environment (Benson et al. 1981 and Stark et al. 1985). Furthermore, the fact
that diesel engines are responsible for a large proportion of the primary organic emissions in
the atmosphere, necessitates measures which help to alleviate not only the nitration associated
with diesels, but also later atmospheric reactions.
In a similar approach to the combustion of primary emissions, the formation of nitro-PAC can
be described in terms of nitration zones, in which nitration is favoured under specific
temperatures and where specific organic species are available for reaction (Figure 6.2).
229
-----71 /I
// I
0 ! NO/N02 /
______ v / /
/
(a)
(b)
Figure 6.2 Two zone nitration concept, such that the free radical combustion generation of nitro-PAC is favoured at high loads and high speeds (a) , whereas electrophilic substitution of surviving PAC is favoured at low loads and low speeds (b).
230
It may be the case, that there are two types of zones which favour the proposed two modes of
nitration. The ftrst zone relates to high temperature combustion generation of nitro-PAC, in
which the greatest PAH radicals and NOx radicals exist (Figure 6.2a) . The second zone
corresponds to lower temperature N02 and HN03 interactions with unburnt fuel, which would
be favoured by higher N02 and PAH survivals (Figure 6.2b). These two types of nitration
zones will be encountered at the extremes of engine power, ie. combustion generation favoured
at high powers compared with N02 & HN03 at low powers.
231
6.2 Conclusions from research 1) The design ofTFSSA minimises the opportunity for artefact formations, and by employing
rapid reproducible sampling a large range of engine conditions may be used on the same day.
The profiling of organic emissions from the combustion chamber over a wide range speeds and
loads is suited to TESSA, especially for trace nitro-PAC emissions.
2) The profiling of the primary PAH and n-alkane emissions showed that low loads result in
the greatest carry over of fuel components to the exhaust stream, whereas the greatest
combustion efficiencies were at high loads. The engine conditions associated with urban
congested driving are those which require the greatest increase in combustion improvements.
The optimum conditions for combustion at low loads would be assigned to a mid-speed of
around 2000 rpm to 2500 rpm. The highest emission concentrations are at high loads and high
speeds, but these are a consequence of the greater fuel consumption not poor combustion.
3) Significant improvements in primary emissions may be achieved by increasing the air
utilization and combustion chamber temperatures in an integrated approach. A strong
correlation with the fuel PAC input and PAC output may enable fuel modifications to be
employed to reduce organics emissions. Greater combustion may give rise to increased
pyrosynthetic/pyrolysis reactions. The conditions for complete combustion are favoured by a
narrow band of high temperatures. A large range of temperatures, termed as a gradient, as
may exist at low load and high speed, enables a range of combustion efficiencies to arise, as
well as specific combustion reactions.
4) The wide variations in the recovery of different organic structures at low load and high
speed, indicate pyrolysis reactions involving alkyl-PAC and pyrolytic cracking of larger n
alkanes into smaller units may be favoured. The much higher recovery of pyrene compared
with naphthalene, fluorene, dibenzothiophene, and phenanthrene at high speed and high load
is strong evidence for combustion generation of pyrene. The possible pyrolysis and cracking
reactions are associated with the lower range of temperatures at low load, whereas the
formation of pyrene is favoured by the extreme temperatures associated with high engine
power.
232
5) Highly mutagenic nitro-PAC are produced during the combustion of standard diesel fuel,
and the formation occurs over a range of speeds and loads. The much lower levels of nitro
PAC found from the combustion chamber sampled using TESSA compared with dilution
tunneUfilter engine sampling, indicates that post-combustion reactions increase nitro-PAH,
some of which may be artefacts formed on the filter.
6) Different speed regimes result in different relationships of the nitro-PAC emissions with
engine load. High speed and high load motoring results in the highest contribution of nitro
PAC to the emissions and also highest output of nitro-PAC per gram of fuel burned. By
contrast, at low speeds, associated with urban areas, the highest nitro-PAC per gram of fuel
consumed is produced at low loads, where the diesel extracts are enriched by P AH surviving
combustion.
7) Nitration can be formed in the engine by two pathways. Each pathway is favoured at a
particular engine regime. Firstly, combustion generation of nitro-PAH is enhanced at high
loads and high speeds where high chamber temperatures favour free radical combustion
reactions forming both nitro-PAC precursors, such as pyrene, and the nitro-PAC. Secondly,
low loads enable nitrogen dioxide and associated nitric acid to nitrate PAC, surviving
combustion late, in the expansion and initial exhaust stages.
8) The close association between the fuel and the nitro-PAC emissions, presents a control
strategy for nitro-PAC emissions based on improving combustion of the primary emissions or
alternatively by lowering the aromatic precursors to nitro-PAC in the fuel. Combustion
generation of nitro-PAC may necessitate post-combustion measures, including innovative
techniques to reduce the important role of the oxides of nitrogen in the nitration reactions.
233
6.3 Future Research Proposals 1) The engine test facilities now allow comparison between different engine specifications, for
example the influence of different fuel injectors on the PAC emissions could be assessed. The
facilities could be used to compare the organic profiles from the up-to-date Prima with that of
the profiles presented in this study. In this way, the degree to which engineering
developments, such as turbocharging and two-stage injectors, reduce organic emissions could
be verified.
2) The nitro-PAC HPLC fractions contains many compounds of that class. Further analysis
of nitro-PAC using gas chromatography with electron capture detection and gas
chromatography/mass spectrometry operated in the negative ion chemical ionisation mode
would enable a greater range of nitro-PAC to be studied. High-performance liquid
chromatography with fluorescence/chemiluminescence detection should be developed to allow
rapid screening of the HPLC fractions. The technique was hindered by the unreliable zinc
reductive columns. Future research would be greatly improved by employing on-line catalytic
reduction of nitro-PAC. Using these techniques and engine sampling with TESSA, further
assessment of the nitration of PAC under high and low engine powers could proceed. The role
of nitrogen containing PAC in the fuel to the formation of nitro-PAC would be of great
interest.
3) Varying of the transfer line between the engine and TESSA would enable the role of post
combustion reactions to be determined. The ultimate design would incorporate a mini-dilution
tunnel with TESSA in series, ie. TESSA as the collector instead of the filter. Such a system
would enable a direct comparison of both forms of engine sampling, and associated artefact
contributions.
4) Running the engine on fuel blends of varied PAH content would enable the effect of the fuel
matrix on nitro-PAH to be determined. Such a study conducted with TESSA would enable the
combustion chamber contribution to the nitro-PAC for different fuels to be established for the
first time.
234
5) Enrichment of PAH components, such as pyrene, of low aromatic fuels, with specific PAH
precursors to nitro-PAC, such as pyrene, would enable the range of nitration products to be
established. Similarly, combustion of radiolabelled precursors to nitro-PAH followed by
isolation of the specific radioactive compounds would provide information on the degree of
nitration induced from one precursor. Such experimentation would require large amounts of
radiolabelled precursor or alternatively combining of consecutive runs.
6) The high mutagenicity that may be attributable to the more polar constituents of the
emissions, requires further research with the aim of detailed characterization and profiling the
effects of combustion over a range of speeds and loads on polar PAC production. The
volatility problems associated with the more polar fractions and the complexity therein,
requires LC detection systems, such as LC/MS. Parallel to the chemical characterisation, a
biological assessment of the identified PAC is essential, if the role of the more hazardous PAC
is to be understood.
235
Literature Cited
Abbass, M.K., Williams, P.T., Andrews, G.E., & Bartle, K.D. (1987) The aging of lubricating oil, the influence of unbumt fuel and particulate SOF contamination. Society of Automotive Engineers. 872085.
Abbass, M.K., Andrews, G.E., Williams, P.T., Bartle, K.D., Davies, I.L., & Tanui, L.K. (1988) Diesel particulate emissions: pyrosynthesis of PAH from hexadecane. Society of Automotive Engineers. 880345.
Abbass, M.K., Andrews, G.E., Kennion, S.J., Williams, P.T., & Bartle, K.D. (1991a) The survivability of diesel fuel components in the organic fraction of particulate emissions from an IDI diesel. Society of Automotive Engineers. 910487.
Abbass, M.K., Andrews, G.E., Ishaq, R.B., Williams, P.T., & Bartle, K.D. (1991b) A comparison of the particulate composition between turbocharged and naturally aspirated Dl diesel engines. Society of Automotive Engineers. 910733.
Aceves, M. & Grimalt, J.O. (1993) Seasonally dependent size distributions of aliphatic and polycyclic aromatic hydrocarbons in urban aerosols from densely populated areas. Environmental Science & Technology. 27, 2896-2908.
Akasaka, Y. & Tamanouchi, M. (1992) Effect of fuel properties on the performance of DI diesel engine with fuel jet impingement. Society of Automotive Engineers. 922213.
Alatas, B., Pinson, J.A., Litzinger, T.A., & Santavicca, D.A. A study of NO and soot evolution in a DI diesel engine via planar imaging. Society of Automotive Engineers. 930973.
Alkidas, A.C. (1984) Relationships between smoke measurements and particulate measurements. Society of Automotive Engineers. 840412.
Amann, C.A., Stivender, D.L., Plee, S.L., & MacDonald, J.S. (1980) Some rudiments of diesel particulate emissions. Society of Automotive Engineers. 800251.
Amann, C.A. & Siegla, D.C. (1982) Diesel particulates- what they are and why. Aerosol Science and Technology. I, 73-101.
Amann, C. A. (1993) Technical options for energy conservation and controlling environmental impact in highway vehicles. International Journal of Vehicle Design. 14 (1), 59-77.
Amstutz, A. & Del Re, L. (1992) EGR feedback control on a turbocharged DI diesel engine. Institute of Mech Engineers. C448/003, 83-88.
Andoh, H. & Shiraishi, K. (1986) Influence on injection and combustion phenomena by elimination of hole nozzle sac volume. Society of Automotive Engineers. 860416.
236
Andrews, G.E., Iheozor-Ejiofor, I.E., Pang, S.W., & Oeapipatanakul, S. (1983) Unburnt hydrocarbon and polynuclear hydrocarbon emissions and their relationship to diesel fuel composition. Institute of Mechanical Engineers. C73/83, 63-74.
Andrews, P.A., Bryant, D., Vitakunas, S., Gouin, M., Anderson, G., McCarry, B.E., Quilliam, M.A., & McCalla, D.R. (1983) Metabolism of nitrated polycyclic aromatic hydrocarbons and formation of DNA-adducts in Salmonella typhimurium. Cooke, M. & Dennis, A.J. (Editors) Polynuclear aromatic hydrocarbons: formation, metabolism and measurement. 89-98.
Andrews, G.E. (1986) Diesel particulates: an introduction to factors that influence their production and occurrence. A Shon Course on Diesel Paniculates. University of Leeds, 22-24 April, Chapter l.
Andrews, G.E., Iheozor-Ejiofor, I.E., & Pang, S.W. (1987) Diesel particulate SOF emissions reduction using an exhaust catalyst. Society of Automotive Engineers. 870251.
Andrews, G.E. & Gibbs, B.M. (1993) Chemical non-catalytic deNO •. Institute of Mechanical Engineers Seminar: Worldwide engine emissions and how to meet them. 137-153.
Aoyagi, Y., Kamimoto, T., Matsui, Y., & Matsuoka, S. (1980) A gas sampling study on the formation processes of soot and NO in a DI diesel engine. Society of Automotive Engineers. 800254.
Arcoumanis, C. (1992) Measurement and prediction of transient NO. emissions in DI engines. Institute of Mechanical Engineers. C448/039, 97-105.
Arcoumanis, C., Chang, I.-C., & Morris, T. (1993) Spray characteristics of single- and twospring diesel fuel injectors. Society of Automotive Engineers. 930922.
Arey, J., Zielinska, B., Atkinson, R., Winer, A.M., Ramdahl, T., & Pitts, J.N., Jr. (1986) The formation of nitro-PAH from the gas-phase reactions of fluoranthene and pyrene with the OH radical in the presence of NO,. Atmospheric Environment. 20 (12), 2239-2345.
Arey, J., Zielinska, B., Atkinson, R., & Winer, A.M. (1987) Polycyclic aromatic hydrocarbon and nitroarene concentrations in ambient air during a wintertime high-NO, episode in the Los Angeles basin. Atmospheric Environment. 21 (6), 1437-1444.
Assaumi, Y., Shintani, M., & Watanabe, Y. (1992) Effects of fuel properties on diesel engine exhaust emission characteristics. Society of Automotive Engineers. 922214
Asou, Y., Fujimoto, H., Senda, J., Tsurutani, K., & Nagae, M. (1992) Combustion in a small DI diesel engine at starting. Society of Automotive Engineers. 920691.
237
Atkinson, R., Arey, J., Zielinska, B., Pitts, J.N., Jr., & Winer, A.M. (1987a) Evidence for the transformation of polycyclic organic matter in the atmosphere. Atmospheric Environment. 21 (10), 2261-2264.
Atkinson, R. & Aschmann, S.M. (1987b) Kinetics of the gas-phase reactions of alkylnaphthalenes with 0 3, NPs and OH radicals at 298±2 K. Atmospheric Environment. 21 (11), 2323-2326.
Atkinson, R., Arey, J., Zielinska, B., & Aschmann, S.M. (1987c) Kinetics and products of the gas-phase reactions of OH radicals and N20 5 with naphthalene and biphenyl. Environmental Science Technology. 21, 1014-1022.
Atkinson, R., Arey, J., Zielinska, B., & Aschmann, S.M. (1990a) Kinetics and nitro-products of the gas-phase OH and N03 radical-initiated reactions of naphthalene-d8 , fluoranthene-d 10,
and pyrene. International Journal of Chemical Kinetics. 22, 999-1014.
Atkinson, R., Tuazon, E. C., & Arey, J. (1990b) Reactions of naphthalene in N20 5-N03-N02-
air mixtures. International Journal of Chemical Kinetics. 22, 1071-1082.
Badger, G.M., Kimber, R.W.L., & Novotny, J. (1964) The formation of aromatic hydrocarbons at high temperatures. XXI: The pyrolysis of n-butlybenzene over a range of temperatures from 300°C to 900°C at 50°C intervals. Aust. J. Chem. 17, 778-786.
Ball, D.J. (1982) Smoke from diesel vehicles- is it time for mass-based emission standard? National Society for Clean Air. 12 (2), 1-4.
Ball, D. (1987) Black lungs and black wall. New Scientist. Feb 12, 28-29.
Barbella, R., Bertoli, C., & Ciajolo, A., & D'Anna, A. (1988) Soot and unbumt liquid hydrocarbon emissions from diesel engines. Combustion Science and Techology. 59, 183-198.
Barbella, R., Ciajolo, A., & D' Anna, A. (1989) Effect of fuel aromaticity on diesel emissions. Combustion and Flame. 17, 267-277.
Bayona, J.M., Markides, K.E., & Lee, M.L. (1988) Characterization of polar polycyclic aromatic compounds in a heavy-duty diesel exhaust particulate by capillary column gas chromatography and high-resolution mass spectrometry. Environmental Science & Technology. 22, 1400-1447.
Bazari, Z. & French, B. A. (1993) Performance and emissions trade-offs for a HSDI diesel engine- an optimization study. Society of Automotive Engineers. 930592
Bechtold, W.E., Dutcher, J.S., Mokler, B. V., Lopez, J.A., Wolf, 1., Li, A.P., Henderson, T.R., & McCiellan, R.O (1984) Chemical and biological properties of diesel exhaust particles collected during selected segments of a simulated driving cycle. Fundamental Applied Toxicology 4, 370-377.
238
Beck, J.N, Uyehara, O.A., Prof. Emeritus, & Johnson, W.P. (1988) Effects of fuel injection on diesel combustion. Society of Automotive Engineers. 880299.
Beland, F.A., Heflich, R.H., Howard, P.C., & Fu, P.P. (1985) The in vitro metabolic activation of nitro polycyclic aromatic hydrocarbons. Harvey. R.G. (Editor). Polycyclic hydrocarbons and carcinogenesis. 371-396.
Belardini, P., Bertoli, C., Del Giacomo, N., & Iorio, B. (1993) Combustion and pollutant emissions from light duty diesel engines: the influence of mixing process and transient operating conditions. The Science of the Total Environment. 134, 285-293.
Benson, J.M., Brooks, A.L., Cheng, Y.S., Henderson, T.R., & White, J.E. (1981) Environmental transformation of 1-nitropyrene on glass surfaces. Bjorseth, A. & Dennis, A.G. (Editors) Polynuclear Aromatic Hydrocarbons. 77-85.
Benson, J.M., Brooks, A.L., Cheng, Y.S., Henderson T.R., & White, J.E. (1984) Cooke, M. & Dennis, A. (Editors) Polynuclear Aromatic Hydrocarbons: Chemistry, characterization and carcinogenesis. 77-85.
Bj0rseth, A.C. (1983) Handbook of Polycyclic Aromatic Hydrocarbons. Marcel Dekker.
Boone, P.M. & Macias, E.S. (1987) Methyl alkanes in atmospheric aerosols. Environmental Science & Technology. 21 (9), 903-909.
Bradow, R.L., Zweidinger, R.B., Black, F.M., & Dietzmann, H.M. (1982) Sampling diesel engine particle and artifacts from nitrogen oxide interaction. Society of Automotive Engineers. 820182.
Braithwaite, A. & Smith, F.J. (1985) Chromatographic Methods. Chapman and Hall.
Brown, K.K. & Poole, C.F. (1984) Determination of 1-nitropyrene by thin-layer chromatography in a diesel exhaust particulate extract. Journal of High Resolution Chromatography & Chromatography Communications. 7, 520-524.
Bulmer, C. (1990) Towards 2000:- how future engines might be developed for the environment. Automotive Engineer. 15 (4), 42-44.
Bundt, J., Herbel, W., Steinhart, H., Franke, S., & Francke, W. (1991) Structure-type separation of diesel fuels by solid phase extraction and identification of the two- and three-ring aromatics by capillary GC-mass spectrometry. Journal of High Resolution Chromatography. 14, 91-97.
Bunting, A. (1992) Electronic diesel fuel injection: the developing science. Automotive Engineer. 17 (6), 38.
239
Burrows, I.E. & Lindsey, A.J. (1961) Formation of polcyclic aromatic hydrocarbons by pyrolysis of simple aliphatic hydrocarbons. Chemistry and Industry. Sept. 2, 1395.
Burtscher, H. ( 1992) Measurement and characteristics of combustion aerosols with special consideration of photoelectric charging and charging by flame ions. Journal of Aerosol Science. 23 (6), 549-595.
Calcote, H.F. ( 1981) Mechanisms of soot nucleation in flames - a critical review. Combustion and Flame. 42, 215-242.
Carnpbell, J., Scholl, J., Hibb1er, F., Bagley, S., Leddy, D., Abata, D., & Johnson, J. (1981) The effect of fuel injection rate and timing on the physical, chemical, and biological character of particulate emissions from a direct injection diesel. Society of Automotive Engineers. 810996.
Campbell, R.M. & Lee, M.L. (1984) Capillary column gas chromatography determination of nitro polycyclic aromatic compounds in particulate extracts. Analytical Chemistry. 56, 1026-1030.
Carroll, J.E. (1991) Autoelectrics - development in engine management systems. Electrotechnology. Oct/Nov, 191-195.
Cartellieri, W. & Tritthart, P. (1984) Particulate analysis of light duty diesel engines (IDI & Dl) with particular reference to the tube oil particulate fraction. Society of Automotive Engineers. 840418.
Castello, G. & Gerbino, T.C. (1993) Analysis of polycyclic aromatic hydrocarbons with an ion-trap mass detector and comparison with other gas chromatographic and high-performance liquid chromatographic techniques. Journal of Chromatography. 642, 351-357.
Chan, T.L. & Gibson, T.L. (1985) Sampling and atmospheric chemistry of particles containing nitrated polycyclic aromatic hydrocarbons. White, C.M. (Editor). Nitrated Polycyclic Aromatic Hydrocarbons. 237-266.
Chang, H.S., Wayte, R., & Spikes, H.A. (1993) Measurement of piston ring and land temperatures in a fired engine using infrared. Tribology Transactions. 36 (1), 104-112.
Chou, M.W. (1984) High performance liquid chromatographic separation of nitro-polycyclic aromatic hydrocarbons and their oxidized derivatives. Cooke, M. & Dennis, A.J. (Editors). Polynuclear Aromatic Hydrocarbons: Chemistry, characterization and carcinogenesis. 145-153.
Choudbury, D.R. (1982) Characterization of polycyclic ketones and quinones in diesel emission particulates by gas chromatography/mass spectrometry. Environmental Science & Technology. 16, 102-106.
240
Ciccioli, P., Brancaleoni, E., Cecinato, A., Di Palo, C., Buttini, P., & Liberti, A. (1986) Fractionation of polar polynuclear aromatic hydrocarbons present in industrial emissions and atmospheric samples and their determination by gas chromatography-mass spectrometry. Journal of Chromatography. 351, 451-464.
Ciccioli, P., Cecinato, A., Cabella, R., & Brancaleoni, E. (1993) The contribution of gasphase reactions to the nitroarene fraction of molecular weight 247 present in carbon particles ofnorthen Italy. Atmospheric Environment. 27A (8) 1261-1270.
Cole, J.A., Bittner, J.D., Longwell, J.P., & Howard, J.B. (1984) Formation mechanisms of aromatic compounds in aliphatic flames. Combustion and Flame. 56, 51-70.
Cooper, B.H., Stanislaus, A., & Hannerup, P.N. (1993) Hydrotreating catalysts for diesel aromatic saturation. Hydrocarbon Processing. June, 83-87.
Cossali, G.E., Coghe, A., & Brunello, G. (1993) Effect of spray-wall interaction on air entrainment in a transient diesel spray. Society of Automotive Engineeers. 930920.
Cowley, L.T., Le Jeune, A., & Lange, W.W. (1993) Effect of fuel composition including aromatics content on emissions from a range of heavy-duty diesel engines. Institute of Mechanical Engineers Seminar: Worldwide engine emission standards and how to meet them. 225-244.
Crittenden, B.D. & Long, R. (1973) Formation of polycyclic aromatics in rich premixed acetylene and ethylene flames. Combustion and Flame. 20, 359-368.
Cuddihy, R.G., Griffith, W.C., & McCiellan, R.O. (1984) Health risks from light-duty diesel vehicles Environmental Science & Technology. 18 (1), 14A-21A.
Cuthbertson, R.D., Shore, P.R., Sundstrom, L., & Heden, P.-0. (1987) Direct analysis of diesel particulate-bound hydrocarbons by gas chromatography with solid sample injection. Society of Automotive Engineers. 870626.
D'Agostino, P.A., Narine, D.R., McCarry, B.E., & Quilliam, M.A. (1983) Clean-up and analysis of nitrated aromatic hydrocarbons in environmental samples. Cooke, M. & Dennis, A.J. (Editors) Polynuclear aromatic hydrocarbons: fonnation, metabolism and measurement. 365-377.
Davies, M.J., Blaikley, D.C.W., Jorgensen, N., Webster, D.E., & Wilkins, A.J.J. (1993) Catalytic control of diesel particulate emissions. Institute of Mechanical Engineers Seminar: Worldwide engine emission standards and how to meet them. 111-121.
Dent, J.C. (1980) Turbulent mixing rate- its effect on smoke and hydrocarbon emissions from diesel engines. Society of Automotive Engineers. 800092.
241
Draper, W.M. (1986) Quantitation of nitro- and dinitro-PAH in diesel exhaust particulate matter. Chemosphere. 15 (4), 437-447.
Draper, W.M., Phillips, J., & Zeller, H.W. (1988) Impact of a barium fuel additive on the mutagenicity and polycyclic aromatic hydrocarbon content of diesel exhaust particulate emissions. Society of Automotive Engineers. 881651.
Duleep, K.G. (1980) Analysis of automotive particulate sampling techniques. Society of Automotive Engineers. 800184.
Diimholz, M., Eifler, G, & Endres, H. (1992) Exhaust-gas re-circulation - a measure to reduce exhaust emissions of DI diesel engines. Society of Automotive Engineers. 920725
El-Shobokoshy, M.S. (1984) The effect of diesel engine load on particulate carbon emission. Atmospheric Environment. 18 (11), 2305-2311.
Escriva, C., Viana, E., Molt6, J.C., Pic6, Y., & Maiies, J. (1994) Comparison of four methods for the determination of polycyclic aromatic hydrocarbons in airborne particulates. Journal of Chromatography. 676, 375-388.
Farrar-Khan, J.R., Andrews, G.E., Jshaq, R., Williams, P.T., & Bartle, K.D. (1993) Quantitiative diesel particulate analysis using gas-chromatography mass-spectrometry. Institute of Mechanical Engineers. 207 (2), 95-106.
Fenn, J.B. (1982) Engines, energy, and entropy. W.H. Freeman & Company.
Ferguson, C.R. (1986) Internal Combustion Engines, Applied Thermosciences. John Wiley & Sons.
Fetzer, J.C. (1989) Gas and liquid-chromatographic techniques. Vo-Dinh, T. (Editor) Chemical analysis of polycyclic aromatic compounds. Chapter 3.
Fillysand, S.A, Kvinge, F., & Betts, W.E. (1993) The influence of diesel fuel properties on particulate emissions in a catalyst equipped European car. Society of Automotive Engineers. 932683.
French, C.C.J. (1987) Power plants for nineties and beyond. Society of Automotive Engineers. 871214.
Fu, F.P., Zhang, Y., Mao, Y.-L., Von Tungeln, L.S., Kim, Y., Jung, H., & Jun, M.-J. (1993) Relationships of structures of nitro-polycyclic aromatic hydrocarbons with highperformance liquid chromatography retention order. Journal of Chromatography. 642, 107-116.
242
Fukuda, M., Tree, D.R., Foster, D.E., & Suhre, B.R. (1992) The effect of fuel aromatic structure and content on direct injection diesel engine particulates. Society of Automotive Engineers. 920110.
Gaddo, P., Settis, M., & Giacomelli, L. (1984) Artifact formation during diesel particulate collection. Cooke, M. & Dennis, A.J. (Editors). Polynuclear Aromatic Hydrocarbons. 437-449.
Galceran, M.T. & Moyano, E. (1993) Determination of oxygenated and nitro-substituted polycyclic aromatic hydrocarbons by HPLC and electrochemical detection. Talanta 40 (5), 615-621.
Galceran, M.T. & Moyano, E. (1994) High-performance liquid chromatography-mass spectrometry (pneumatically assisted electrospray) of hydroxy polycyclic aromatic hydrocarbons. Journal of Chromatography. 683, 9-19.
Gachanja, A.N. (1993) Analysis of polycyclic aromatic hydrocarbons by liquid chromatography. Chromatography and Analysis. Feb/March, 5-7.
Gachanja, A. N. (1994) University of Plymouth, Personal Communication.
Gibson, T.L., Ricci, A. I., & Williams, R.L. (1981) Measurement of polynuclear aromatic hydrocarbons, their derivatives & their reactivity in diesel automobile exhaust. Bjorseth, A. & Dennis, A.G. (Editors) Polynuclear Aromatic Hydrocarbons. Battelle Press, 707-717.
Gibson, T.L. (1982) Nitro derivatives of polynuclear aromatic hydrocarbons in airborne and source particulate matter. Atmospheric Environment. 16 (8) 2037-2041.
Gomes, P.C.F & Yates, D.A. (1992) The influence of some engine operating parameters on particulate emissions. Society of Automotive Engineers. 922222.
Gonaza.Iez, M.A., Rodriguez, G.B., Galiasso, R., & Rodriguez, E. (1993) A low emission diesel fuel: hydrocracking production, characterization and engine evaluations. Society of Automotive Engineers. 932731.
Gorse, R.A., Jr., Riley, T.L., Ferris, F.C., Pero, A.M. & Skewes, L.M. (1983) 1-Nitropyrene concentration and bacterial mutagenicity in on-road particulate emissions. Environmental Science & Technology. 17, 198-202.
Grasso, P., Mann, A., & lrvine, D. (1988) Review of literature on the possible carcinogenicity of automotive emissions. Final Report. Robens Institute, University of Surrey.
Greeves, G. (1979) Response of diesel combustion systems to increase of fuel injection rate. Society of Automotive Engineers. 790037.
243
Grosjean, D., Fung, K., & Harrison, J. (1983) Interactions of polycyclic aromatic hydrocarbons with atmospheric pollutants Environmental Science & Technology. 17, 673-679.
Hamilton, R.S. & Mansfield, T.A. (1993) The soiling of materials in the ambient atmosphere. Atmospheric Environment. 274 (8), 1369-1374.
Haddad, S.D. (1984) Combustion and heat release in diesel engines. Haddad, S.D. & Watson, N. (Editors) Principles and performance in diesel engineering. Ellis Horwood, Chapter 1.
Hammerle, R.H., Shiller, J.W., & Schwarz, M.J. (1991) Global Climate Change. Transactions of Engineering for gas turbines & power. 113, 448-455.
Handa, T., Yamauchi, T., Sawai, K., Yamamura, T., Koseki, Y., & Ishii, T. (1984) In situ emission levels of carcinogenic and mutagenic compounds from diesel and gasoline engine vehicles on an expressway. Environmental Science & Technology. 18 (2), 895-202.
Hansen, A.M., Olsen, I.L.B., Holst, E., & Poulsen, O.M. (1991) Validation of a highperformance liquid chromatography/fluorescence detection method for the simultaneous quantification of fifteen polycyclic aromatic hydrocarbons. Ann. Occup. Hyg. 35 (6), 603-611.
Hanson, R.L., Henderson, T.R., Hobbs, C.H., Clark, C.R., Carpenter, R.L., & Dutcher, J.S. (1983) Detection of nitroaromatics compounds on coal combustion particles. Journal of Toxicology and Environmental Health. 11 971-980.
Harris, G.W., Mackay, G.l., Iguchi, T., Schiff, H.I., & Schuetzle, D. (1987) Measurement of N02 and HN03 in diesel exhaust by tunable diode laser absorption spectrometry. Environmental Science & Technology. 21, 299-304.
Hartung, A., Kraft, J., Schulze, J, KieJ3, H., & Lies, K.-H. (1984) The identification of nitrated polycyclic aromatic hydrocarbons in diesel particulate extracts and their potential formation as artifacts during particulate collection. Chromatographia. 19, 269-273.
Haupais, A. (1982) A phenomenological approach of combustion modelling in DI diesel engines. Institute of Mechanical Engineers. C128/82, 229-236.
Hayakawa, K., Butoh, M, & Miyazaki, M (1992) Determination of dinitro- and nitropyrenes in emission particulates from diesel and gasoline engine vehicles by liquid chromatography with chemiluminescence detection after precolumn reduction. Analytica Chimica Acta. 266, 251-256.
Hayona, S., Jang-Ho, L., Furuya, K., Kikuchi, T., Someya, T., Oikawa, C., Iida, Y., Matsushita, H., Kinouchi, T., Manabe, Y., & Ohnishi, Y. (1985) Formation of hazardous substances and mutagenicity of P AH produced during the combustion process in a diesel engine. Atmopsheric Environment. 19 (6), 1009-1015.
244
Helmig, D., Arey, J., Atkinson, R., Harger, W.P., & McEiroy, P.A. (1992) Products of the OH radical-initiated gas-phase reaction of fluorene in the presence of NO,. Atmospheric Environment. 26A (9), 1735-1745.
Henderson, T.R., Royer, R.E., Clark, C.R., Harvey, T.M., & Hunt, D.F. (1982) MS/MS analysis of diesel emissions and fuels treated with N02• Journal of Appled Toxicology 2 (5), 231-237.
Henderson, T.R., Sun, J.D., Royer, R.E., Clark, C.R., Li, A.P, Harvey, T.M., Hunt, D.F., Fulford, J.E., Lovette, A.M., & Davidson, W.R. (1983) Triple-quadrupole mass spectrometry studies of nitroaromatic emissions from different diesel engines. Environmental Science & Technology. 17 (8), 443-449.
Henderson, T.R., Sun, J.D., Albert, L.P., Hanson, R.L., Bechtold, W.E., Harvey, T.M., Shabanowitz, J., & Hunt, D.F. (1984) GC/MS and MS/MS studies of diesel exhaust mutagenicity and emissions from chemically defined fuels. Environmental Science & Technology. 18 (6), 428-429.
Henein, N.A. (1973) Diesel engines combustion and emissions. Springer, G.S. & Patterson, D.J. (Editors) Engine emissions. Pollutant formation and measurement. Plenum Press. New York. 211- 266.
Henk, M.G., Williamson, W.B., & Silver, R.G. (1992) Diesel catalysts for low particulates and low sulfate emissions. Society of Automotive Engineers. 920368.
Herr, J.D., Dukovich, M., Letz, S.S., Yergey, J.A., Risby, T.H., & Tejada, S.B. (1982) The role of nitrogen in the observed direct mircobial mutagenic activity for diesel engine combustion in a single-cylinder DI engine. Society of Automotive Engineers. 820467.
Hillier, V.A.W. (1991) Fundamentals of Motor Vehicle Technology. Stanley Thomes Ltd. England, Chapter 29.
Hirakouchi, N., Fukano, I., & Nagano, H. (1990) Measurement of unregulated exhaust emissions from heavy duty diesel engines with mini-dilution tunnel. Society of Automotive Engineers. 900643.
Hisamatsu, Y., Nishimura, T., Tanabe, K., & Matsushita, H. (1986) Mutagenicity of the photochemical reaction products of pyrene with nitrogen dioxide. Mutation Research. 172, 19-27.
Hites, R.A. (1989) Mass spectrometry of polycyclic aromatic compounds. Vo-Dinh, T. (Editor) Chemical analysis of polycyclic aromatic compounds. Chapter 8.
245
Hochhauser, A.M., Gorse, R.A., Jr., Reuter, R.M., Benson, J.D., Koehl, W.J., Rutherford, J.A., Bums, V.R., & Painter, L.J. (1992) Speciation and calculated reactivity of automotive exhaust emissions and thier relation to fuel properties - Auto/oil air quality improvement research program. Society of Automotive Engineers. 920325.
Horrocks, R.W. (1993) Light duty diesels- an update on the emissions challenge. Institute of Mechanical Engineers Seminar: Worldwide engine emissions and how to meet them. 155-169.
Howard, J.B. & Longwell, J.P. (1983) Formation mechanisms of PAH and soot in flames. Cooke, M. & Dennis, A.J. (Editors) Polynuclear aromatic hydrocarbons: formation, metabolism and measurement. 27-62.
International Agency for Research on Cancer (IARC) (1989) Monographs on the evaluation of carcinogenic risk to humans. Diesel and gasoline engine exhausts and some nitroarenes. WHO, Lyon, France, 46.
lida, N., Suzuki, Y., Sato, G.T., & Sawada, T. (1986) Effects of intake oxygen concentration on the characteristics of particulate emissions from a D.I. diesel engine. Society of Automotive Engineers. 861233.
lida, N. & Sato, G.T. (1988) Temperature & mixing effects on NO, and particulate. Society of Automotive Engineers. 880424.
Imaizumi, N., Hayakawa, K., Miyazaki, M., Imai, K. (1989) Stability of bis(2,4,6-trichlorophenyl) oxalate in high-performance liquid chromatography for chemiluminscence detection. Analyst. 114, 161-164.
Imaizumi, N., Hayakawa, K., Suzuki, Y., & Miyazaki, M. (1990) Determination of nitrated pyrenes and their derivatives by high performance liquid chromatography with chemiluminescence detection after online electrochemical reduction. Biomedical Chromatography. 4 (3), 108-112.
Jliger, J. (1978) Detection and characterization of nitro derivatives of some polycyclic aromatic hydrocarbons by fluorescence quenching after thin-layer chromatography: application to air pollution analysis. Journal of Chromatography. 152, 575-578.
Jensen, T.E. & Hites, R.A. (1983) Aromatic diesel emissions as a function of engine conditions. Analytical Chemistry. 55 (4), 594-599.
Jensen, T.E., Richert, J.F.O, Cleary, A.C., LaCourse, D.L., & Gorse, R.A., Jr. (1986) 1-Nitropyrene in used engine oil. Journal of the Air Pollution Control Association. 36 (11), 1255-1257.
Jin, Z. & Rappaport, S.M. (1983) Microbore liquid chromatography with electrochemical detection for determination of nitro-substituted polynuclear aromatic hydrocarbons in diesel soot. Analytical Chemistry. 55, 1778-1781.
246
Kamimoto, T. & Yagita, M. (1989) Particulate formation and flame structure in diesel engines. Society of Automotive Engineers. 890436
Karnimoto, T. & Kobayashi, H. (1991) Combustion processes in diesel engines. Prog. Energy Combust. Sci. 17, 163-189.
Karcher, W., Fordham, R.J., Dubois, J.J., Glaude, P.G.J.M., & Ligthart, J.A.M. (1985) Spectral atlas of polycyclic aromatic compounds. Vol. I, D.Reidel.
Karimi, E.R. (1989) High-speed photography of fuel spray and combustion events in a production diesel engine and combustion bomb. Institute of Mechanical Engineers. 203, 269-281.
Kelly, G.W., Bartle, K.D., Clifford, A.A., & Robinson, R.E. (1992) Application of coupled LC-GC to the analysis of the polar fraction of diesel particulate matter. Journal of High Resolution Chromatography. 15, 526-530.
Kerswill, J. (1992) Diesel: pump it up. The Guardian. 191h October, 12.
Kingston, S.T. (1994) Genotoxicology of diesel engine exhaust emissions in cultered mammalian cells. Thesis, University of Plymouth, June.
Kittelson, D. B., Du, C.-J., & Zweidinger, R.B. (1984) Measurements of polycyclic aromatic compounds in the cylinder of an operating diesel engine. Society of Automotive Engineers. 840364.
Kobayashi, S. & Imai, K. (1980) Determination of fluorescent compounds by high performance liquid chromatography with chemiluminscence detection. Analytical Chemistry. 52, 424-427.
Kobayashi, S., Sakai, T., Nakahira, T., Komori, M., & Tsujimura, K. (1992) Measurement of flame temperature distribution in D.I. diesel engine with high pressure fuel injection. Society of Automotive Engineers. 920692.
Konig, J., Balfanz, E., Funcke, W., & Romanowski, T. (1983) Determination of oxygenated polycyclic aromatic hydrocarbons in airborne particulate matter by capillary gas chromatography and gas chromatography/mass spectrometry. Analytical Chemistry. 55, 599-603.
Konno, M., Chikahisa, T., Murayama, T., & Iwamoto, M. (1992a) Catalytic reduction of NOx in actual diesel engine exhuast. Society of Automotive Engineers. 920091.
Konno, M., Chikahisa, T., & Murayarna, T. (1992b) Reduction of smoke and NOx by strong turbulence generated during the combustion process in D.I. diesel engines. Society of Automotive Engineers. 920467.
247
Konno, M., Chikahisa, T., & Murayama, T. (1993) An investigation on the simultaneous reduction of particulate and NOx by controlling both the turbe1ence and the mixture formation in DI diesel engines. Society of Automotive Engineers. 932797.
Kowalewicz, A. ( 1984) Combustion systems of high-speed piston I. C. engines. Elsevier
Kraft, J. & Lies, K.-H. (1981) Polycyclic aromatic hydrocarbons in the exhaust of gasoline and diesel vehicles. Society of Automotive Engineers. 810082.
Kruzel, E.L., Lafleur, A.L., Braun, A.G., Longwell, J.P., Thilly, W.G., & Peters, W.A. ( 1991) A high-yield sampler for toxicological characterization of complex mixtures in combustion effluents. Environmental Health Perspectives. 90, 305-314.
Kuhler, M., Kraft, J., Bess, H., Heeren, U., & Schilrmann, D. (1994) Comparison between measured and calculated concentrations of nitrogen oxides and ozone in the vicinity of a motorway. The Science of the Total Environment. 146/147, 387-394.
Kwakman, P.J.M. & Brinkman, U.A.Th. (1992) Peroxyoxalate chemiluminescence detection in liquid chromatography. Analytica Chimica Acta. 266, 175-192.
Lach, G. & Winckler, J. (1988) Specific problems of sampling and measuring diesel exhaust emissions. Society of Automotive Engineers. 881763.
LaCourse, P.L. & Jensen, T.E. (1986) Determination of 1-nitropyrene in extracts of vehicle particulate emissions. Analytical Chemistry. 58 (8), 1894-1895.
Ladommatos, N., Horrocks, R., & Cooper, L. (1993) Control of the start of combustion in a direct injection diesel engine using an optical sensor. Institute of Mechanical Engineers Seminar: Worldwide engine emission standards and how to meet them. 183-196.
Laity, J.L., Malbin, M.D., Haskwell, W.W., & Doty, W.I. (1973) Mechanisms of polynuclear aromatic hydrocarbon emissions from automotive engines. Society of Automotive Engineers. 730835.
Langley, J. (1991) What's clean and lean and a godsend to a green? The Daily Telegraph. 30th January, Driving section, l.
Later, D.W., Wilson, B.W., & Lee, M.L. (1985) Standardization of alumina and silica adsorbents used for chemical class separations of polycyclic aromatic compounds. Analytical Chemistry. 57, 2979-2984.
Lee, M.L., Novontny, M. V., & Bartle, K.D. (1981) Analytical Chemistry of Polycyclic Aromatic Compounds. Academic Press.
Lenner, M. (1987) Nitrogen dioxide in exhaust emissions from motor vehicles. Atmospheric Environment. 21 (1), 37-43.
248
--------------
Levin, J.-0., Nilsson, C.-A., & Norstrom, A (1984) Sampling and analysis of polycyclic aromatic hydrocarbons (PAH) from two-stroke chain-saw engines. Chemosphere. 13 (3), 427-435.
Levine, S.P. & Skewes, L.M. (1982) High-performance semi-preparative liquid chromatography of diesel engine emission particulate extracts. Journal of Chromatography. 235, 532-535.
Li, C.-H. (1982) Piston thermal deformation and friction considerations. Society of Automotive Engineers. 820086.
Li, H. & Westerholm, R. (1994) Determination of mono- & di-nitro polycyclic aromatic hydrocarbons by on-line reduction and high performance liquid chromatography with chemiluminescence detection. Journal of Chromatography. 664, 177-182.
Liberti, A., Ciccioli, P., Cecinato, A., Brancaleoni, E. & Di Palo, C. (1984) Determination of nitrated-polyaromatic hydrocarbons (nitro-PAHs) in environmental samples by high resolution chromatographic techniques. Journal of High Resolution Chromatography & Chromatography Communications. 7, 389-397.
Lilly, L.C.R. (1984) Diesel Engine Reference Book. Butterworths.
Linder, W., Posch, W., Wolfbeis, O.S., & Tritthart, P. (1985) Analysis of nitro-PAHs in diesel exhaust particulate extracts with multi-column HPLC. Chromatographia. 20 (14), 213-218.
Lindskog, A. (1983) Transformations of polcyclic aromatic hydrocarbons during sampling. Environmental Health Perspectives, 47, 81-84.
Liu, T.-Y. & Robbat, A., Jr. (1991) High-performance liquid chromatography retention index and detection ofnitrated polycyclic aromatic hydrocarbons. Journal of Chromatography. 539, 1-14.
MacCrehan, W.A., May, W.E., Yang, S.D., & Benner, B.A., Jr. (1988) Determination of nitro polynuclear aromatic hydrocarbons in air and diesel particulate using liquid chromatography with electrochemical and fluorescence detection. Analytical Chemistry. 60, 194-199.
Maher, W., Pellegrino, F., & Furlonger, J. (1989) Determination of polycylic aromatic hydrocarbons in air particulate samples. Microchemical Journal. 39, 160-165.
Martin, B. & Bigeard, P.H. (1992) Hydrotreatment of diesel fuels- Its impact on light-duty diesel engine pollutants. Society of Automotive Engineers. 922268.
Matsui, Y. & Sugihara, K. (1986) Sources of hydrocarbon emissions from a small direct injection diesel engine. Japananese Society of Automotive Engineering Review. 7 (3), 4-11.
249
Matsuoka, S. (1990) Combustion in the diesel engine. Weaving, J.H. (Editor) Internal Combustion Engineering Science & Technology. Elsevier Applied Science, 333-384.
Mayer, W.J., Lechman, D.C., & Hilden, D.L. (1980) The contribution of engine oil to diesel exhaust particulate emissions. Society of Automotive Engineers. 800256.
May, W.E. & Wise, S.A. (1984) Liquid chromatographic determination of polycyclic aromatic hydrocarbons in air particu1ate extracts. Analytical Chemistry. 56 (2), 225-232.
Melikian, A.A., Lavoie, E.J., Hecht, S.S., & Hoffmann, D. (1983) On the enhancing effect of a bay-region methyl group in 5-methylchrsyene carciongenesis. Cooke, M. & Dennis, A.J. (Editors). Polynuclear Aromatic Hydrocarbons: Formation, Metabolism, and Measurement. Battelle Press. 861-875
Mermelstein, R., Kiriazides, D., Butler, M., McCoy, E.C., & Rosenkranz, H.S. (1981) The extraordinary mutagenicity of nitropyrenes in bacteria. Mutation Research. 89, 187-198.
Mills, G.A., Howarth, J.S., & Howard, A. G. (1984) The effect of diesel fuel aromaticity on polynuclear aromatic hydrocarbon exhaust emissions. Journal of the Institute of Energy. March, 273-286.
Miyamoto, N., Ogawa, H., Shibuya, M., & Suda, T. (1992) Description of diesel emissions by individual fuel properties. Society of Automotive Engineers. 922221.
Monaghan, M.L. (1981) High speed direct injection diesel for passenger cars. Society of Automotive Engineers. 810477.
Montreuil, C.N., Ball, J.C., Gorse, R.A., Jr., & Young, W.C. (1992) Solvent extraction efficiencies of mutagenic components from diesel particles. Mutation Research. 282, 89-92.
Muramatsu, G., Abe, A., Furuyama, M., & Yoshida, K. (1993) Catalytic reduction of NO, in diesel exhaust. Society of Automotive Engineers. 930135
Murayama, T., Miyamoto, N., Chikahisa, T., & Yamane, K. (1986) Effects of combustion and injection systems on unburnt HC and particulate emissions from a DI diesel engine. Society of Automotive Engineers. 861232.
Naber, D., Lange, W.W., Reglitzky, A.A., Schiifer, A., Gairing, M., & Le Jeune, A. (1993) The influence of fuel properties on exhaust emissions from advanced Mercedes Benz diesel engines. Society of Automotive Engineers. 932685.
Nahum, A. (1989) Direct approach. Autocar & Motor. 10 May.
250
Nakagawa, R., Kitamori, S., Horikawa, K., Nakashima, K., & Tokiwa, H. (1983) Identification of dinitropyrenes in diesel-exhaust particles. Their probable presence as major mutagens. Mutation Research. 124, 201-211.
Nelson, P.F. (1989) Combustion-generated polycyclic aromatic hydrocarbons in diesel exhaust emissions. Fuel. 68, 283-286.
Newton, D.L., Erickson, M.D., Tomer, K.B., Pellizzari, E.D., Gentry, P., & Zweidinger, R.B. (1982) Identification of nitroaromatics in diesel exhaust particulate using gas chromatography/negative ion chemical ionization mass spectrometry and other techniques. Environmental Science & Technology. 16, 206-213.
Nielsen, T. (1983) Isolation of polycyclic aromatic hydrocarbons and nitro derivatives in complex mixtures. Analytical Chemistry. 55, 286-290.
Nielsen, T., Ramdahl, T., & Bj0Tseth, A. (1983) The fate of airborne polcyclic organic matter. Environmental Health Perspectives. 47, 103-114.
Nielsen, T. (1984) Reactivity of polycylic aromatic hydrocarbons towards nitrating species. Environmental Science & Technology. 18, 157-163.
Nikanjam, M. (1993) Development of the first CARB certified California alternative diesel fuel. Society of Automotive Engineers. 930728.
Niles, R. & Tan, Y.L. (1989) Determination of polynuclear aromatic hydrocarbons and mononitrated derivatives in air and diesel particulates. Analytica Chimica Acta. 221, 53-63.
Ning, M., Yuan-Xian, Z., Zhen-Huan, S., & Hu-Guo-dong (1991) Soot formation, oxidation and its mechanism in different combustion systems and smoke emission pattern in DI diesel engines. Society of Automotive Engineers. 910230.
Noble, D. (1994) Perkins Technology, Peterborough, Personal communication.
Obuchi, A., Aoyama, H., Ohi, A., & Ohuchi, H. (1984) Determination of polycyclic aromatic hydrocarbons in diesel exhaust particulate matter and diesel fuel oil. Journal of Chromatography. 312, 247-259.
Oehme, M., MaruJ, S., & Stray, H. (1982) Determination of nitrated polycyclic hydrocarbons in aerosols using capillary gas chromatography combined with different electron capture detection methods. Journal of High Resolution Chromatography & Chromatography Communications. 5, 417-423.
Oehme, M. (1985) Negative ion chemical ionization mass spectrometry - a useful technique for the selective detection of polar substituted polycyclic aromatic hydrocarbons with mutagenic properties. Chemosphere. 14 (9), 1285-1297.
251
Ohgaki, H., Negisk:i, C., Wakabayashi, K., Kusama, K., Sato, S., & Sugimura, T. (1984) Induction of sarcomas in rats by subcutaneous injection of dinitropyrenes. Carcinogenesis. 5 (5), 583-585.
Ohnishi, Y., Kachi, K., Sato, K., Tahara, 1., Takeyoshi, H., & Tokiwa, H. (1980) Detection of mutagenic activity in automobile exhaust. Mutation Research. 77, 229-240.
Paputa-Peck, M.C, Marano, R.S., Schuetzle, D., Riley, T.L., Hampton, C. V., Prater, T.J., Skewes, L.M., Jensen, T.E., Ruehle, P.H., Bosch, L.C., & Duncan, W.P. (1983) Determination of nitrated polynuclear aromatic hydrocarbons in particulate extracts by capillary column gas chromatography with nitrogen selective detection. Analytical Chemistry. 55, 1946-1954.
Parliamentary Office of Science and Technology (1994) Breathing in our cities - urban air pollution and respiratory health. Report.
Pemberton (1995) University of Plymouth, Personal communication.
Petch, G.S., Trier, C.J., Rhead, M.M., Fussey, D.E. & Millward, G.E. (1987) The development of a novel exhaust sampling technique with particular relevance to polycyclic aromatic hydrocarbons. Institute of Mechanical Engineers. C340/87, 97-105.
Pederson, T.C. & Siak, J.-S. (1981) The role of nitroaromatic compounds in the direct-acting mutagenicity of diesel particle extracts. Journal of Applied Toxicology. 1 (2), 54-60.
Ffeffer, H.-U. (1994) Ambient air concentrations of pollutants at traffic-related sites in urban areas of North Rhine-Westphalia, Germany. The Science of the Total Environment. 146/147, 263-273.
Pipho, M.J., Kittelson, D.B., & Zarling, D.D. (1991) N02 formation in a diesel engine. Society of Automotive Engineers. 910231.
Pitts, J.N., Jr., Van Cauwenberghe, K.A., Grosjean, D., Schmid, J.P., Fitz, D.R., Be1ser, W.L., Jr., Knudson, G.B., & Hynds, P.M. (1978) Atmospheric reactions of polycyclic aromatic hydrocarbons: facile formation of mutagenic nitro derivatives. Science. 202, 515-519.
Pitts, J.N., Jr. (1983) Formation and fate of gaseous and particulate mutagens and carcinogens in real and simulated atmospheres. Environmental Health Perspectives. 47, 115-140.
Pitts, J.N., Jr., Arey, J., Zeilinska, B., Winer, A.M., & Atkinson, R. (1985) Determination of 2-nitrofluoranthene and 2-nitropyrene in ambient particulate organic matter: evidence for atmospheric reactions. Atmospheric Environment. 19 (10), 1601-1608.
Plee, S.L. & MacDonald, J .S. (1980) Some mechanisms affecting the mass of diesel exhaust particulate following a dilution process. Society of Automotive Engineers. 8001986.
252
Poole, C.F. (1982) The electron-capture detector in capillary gas chromatography. Journal of HRC & CC. 5, 454-471.
Pou1ton, M.L. (1994) Alternative engim!s for road vehicles. Computational Mechanics Publications.
Quality of Urban Air Review Group (QUARG) (1993). Diesel vehicle emissions and urban air quality. Second Report, December.
Ramdahl, T. & Urdal, K. (1982) Determination of nitrated polycyclic aromatic hydrocarbons by fused silica capillary gas chromatography/negative ion chemical ionization mass spectrometry. Analytical Chemistry. 54, 2256-2260.
Ramdahl, T. (1983) Polycyclic aromatic ketones in environmental samples. Environmental Science & Technology. 17, 666-670.
Ramdahl, T., Sweetman, J.A., Zielinska, B., Atkinson, R., Winer, A.M., & Pitts, J.N., Jr. (1985) Analysis of mononitro-isomers of fluoranthene and pyrene by high resolution capillary gas chromatography/mass spectrometry. Journal of High Resolution Chromatography & Chromatography Communications. 8, 849-853.
Rao, K.K., Winterbone, D.E., & Clough, E. (1993) Influence of swirl on high pressure injection in hydra diesel engine. Society of Automotive Engim!ers. 930978.
Rappaport, S.M., Jin, Z.L., & Xu, X.B. (1982) High-performance liquid chromatography with reductive electrochemical detection of mutagenic nitro-substituted polynuclear aromatic hydrocarbons in diesel exhausts. Journal of Chromatography. 240, 145-154.
Reece, C.E. & Scott, P.W. (1987) Quantitative analysis by gas chromatography. In: Katz, E. (Editor). Quantitative analysis using chromatographic techniques. 157-191.
Rhead, M.M., Fussey, D.E., Millward, G. E., & Trier, C.J. (1991) Pathways of PAH radiotracers during transient testing of a modem HSDI diesel engine. Final report, SERC grant, GR/E/76780.
Rhead, M.M. & Trier, C.J. (1992) Fuel residues and organic combustion products in diesel exhaust emissions: sources, sampling and analysis. Trends in Analytical Chemistry. 11 (7), 255-259.
Richards, L.W., Bergstrom, R.W., Jr., & Ackerman, T.P. (1986) The optical effects of fineparticle carbon on urban atmosphere. Atmospheric Environment. 20 (2), 387-396.
Risby, T.H. & Lestz, S.S. (1983) Is the direct mutagenic activity of diesel particulate matter a sampling artefact? Environmental Science & Technology. 17 (10), 621-624.
253
Rivers, K.J., Paassen, C.W.C., Booth, M., & Marriott, J.M. (1993) Future diesel fuel quality - balancing requirements. Institute of Mechanical Engineers Seminar: Worldwide engine emission standards and how to meet them. 209-223.
Robards, K. & Worsfold, P.J. (1992) Analytical applications of liquid-phase chemiluminescence. Analytica Chimica Acta. 266, 147-173.
Robbat, A., Jr., Corso, N.P., Doherty, P.J., & Wolf, M.H. (1986) Gas chromatographic chemiluminescent detection and evaluation of predictive models for identifying nitrated polycyclic aromatic hydrocarbons in a diesel fuel particu1ate extract. Analytical Chemistry. 58, 2078-2084.
Robbins, W.K. & McElroy, F.C. (1982) Rational selection of sample preparation techniques for the measurement of polynuclear aromatics. Cooke, M., Dennis, A.J., & Fisher, G.L. (Editors) PAR's Physical and biological Chemistry. 673-686.
Roj, A. (1992) Environmentally friendly diesel fuels -The Swedish experience. 11 lh European Automotive Symposium, Sorrento, November 19-20.
Rosenkranz, H.S. & Mermelstein, R. (1983). Mutagenicity and genotoxicity. All nitrocontaining chemicals were not created. Mutation Research. 114, 217-267.
Ross, D.S., Hum, G.P., & Schmitt, R.J. (1988) Nitration of pyrene by N02 and N20 4 • In Elbert, L.B. (Editor) Polynuclear aromatic compounds. American Chemical Society, Chapter 9.
Rubey, W.A., Hall, D.L., Torres, J.L., Dellinger, B., & Carnes, R.A. (1983) The thermal degradation behaviour of selected polynuclear aromatic hydrocarbons. Cooke, M. & Dennis, A.J. {Editors). Polynuclear Aromatic Hydrocarbons: Formation, Metabolism, and Measurement. Battelle Press. 861-875
Rudolf, W. (1994) Concentration of air pollutants inside cars driving on highways and in downtown areas. The Science ofthe Total Environment. 1461147, 433-444.
Ryan, T.W., Ill, Stroment, J.O., Wright, B.R., & Waytulonis, R. (1981) The effects of fuel properties and composition on diesel engine exhaust emissions-a review. Society of Automotive Engineers. 810953.
Ryan, T. W., Ill & Erwi.n, J. (1993) Diesel fuel composition effects on ignition and emissions. Society of Automotive Engineers. 932735.
Saito, T., Tokura, N., & Katoh, T. (1982) Analysis of factors affecting the formation of major mutagenic substances in diesel particulate extracts. Society of Automotive Engineers. 821244.
Saito, T., Daisho, Y., Uchida, N., & Ikeya, N. (1986) Effects of combustion chamber geometry on diesel combustion. Society of Automotive Engineers. 861186.
254
Salmeen, I., Durisin, A.M., Prater, T.J., Riley, T., & Schuetzle, D. (1982) Contribution of 1-nitropyrene to direct-acting Ames assay mutagencities of diesel particulate extracts. Mutation Research. 104, 17-23.
Sasaki, Y., Endo, R. & Koido, Y. (1980) Direct mutagens in the gaseous component of automobile exhaust detected with Bacillus subtillis spores. Mutation Research. 19, 181-184.
Scheepers, P.T.J. & Bos, R.P. (1992a) Combustion of diesel fuel from a toxicological perspective. 11. Toxicity. lnt. Arch. Occup. Environ. Health. 64, 163-77.
Scheepers, P.T.J. & Bos, R.P. (1992b) Combustion of diesel fuel from a toxicological perspective. I. Origin of incomplete combustion products. lnt. Arch. Occup. Environ. Health. 64, 149-161.
Scheepers, P.T.J., Velders, D.D., Martens, M.H.J., Noordhoek, J., Bos, R.P. (1994) Gas chromatographic-mass spectrometric determination of nitro polycyclic aromatic hydrocarbons in airborne particulate matter from workplace atmospheres contaminated with diesel exhaust. Journal of Chromatography. 677, 107-121.
Schenker, M.B. (1980) Diesel exhaust- an occupational carcinogen? Journal of Occupational Medicine. 22 (1}, 41-46.
Schindler, K.-P. (1992) Integrated diesel european action (IDEA): study of diesel combustion. Society of Automotive Engineers. 920591.
Schmeltz, I. & Hoffman, D. (1976) Formation of polynuclear hydrocarbons from combustion of organic matter. Fredenthal, R.I. & Jones, P.W. (Editors) Polynuclear aromatic hydrocarbons: Chemistry, metabolism, and carcinogenesis. Vol. 1, Raven Press, 225-239.
Schneider, E., Krenmayr, P., & Varmuza, K. (1990) A routine method for the analysis of the mononitro-PAH in immission and emission samples. Monatshefte fii.r Chemie Chemical Monthly. 121, 393-401.
Schuetzle, D., Lee, F.S-C., Prater, T.J., Tejada, S.B. (1981) The identificaton of polynuclear aromatic hydrocarbon (PAH) derivatives in mutagenic fractions of diesel particulate extracts. Intern. J. Environ. Anal. Chem. 9, 93-144.
Schuetzle, T.L., Riley, T.L., Prater, T.J., Harvey, T.M., & Hunt, D. F. (1982) Analysis of nitrated polycyclic aromatic hydrocarbons in diesel particulates. Analytical Chemistry. 54, 265-271.
Schuetzle, D. (1983) Sampling of vehicle emissions for chemical analysis and biological testing. Environmental Health Perspectives. 47, 65-80.
255
Schuetzle, D. & Perez, J.M. (1983) Factors influencing the emissions of nitrated-polynuclear aromatic hydrocarbons (nitro-PAH) from diesel engines. Journal of Air Pollution Control Association. 33 (8). 751-755.
Schuetzle, D. & Frazier, J.A. (1986) Factors influencing the emission of vapor and particulate phase components from diesel engines. Ishinishi, N., Koizumi, A., McClellan, R.O., & Stober, W. (Editors) Carcinogenic and mutagenic effects of diesel engine exhaust. Elsevier, Amsterdam, 41-63.
Schulze, J., Hartung, A., Kie6, H., Kraft, J., & Lies, K.-H. (1984) Identification of oxygenated polycyclic aromatic hydrocarbons in diesel particulate matter by capillary gas chromatography and capillary gas chromatography/mass spectrometry. Chromatographia. 19, 391-397.
Seaton, A., MacNee, W., Donaldson, K., & Godden, D. (1995) Particulate air pollution & acute health effects. The Lancet. 345, 176-178.
Sellestrom, U., Jansson, B., Bergman, A, & Alsberg. T. (1987) Selective and sensitive analysis of nitro-PAH. Chemosphere. 16 (5), 945-952.
Serageldin, M. A. (1981) Soot formation in small flames. Society of Automotive Engineers. 811197.
Shakal, J. & Martin, I .K. (1990) Effects of auxiliary injection on diesel engine combustion. Society of Automotive Engineers. 900398.
Shay, E.G. (1993) Diesel fuel from vegetable oils: status and opportunities. Biomass and Bioenergy. 4 (4), 227-242.
Shioji, M., Kimoto, T., Okamoto, M., & Ikegami, M. (1989) An anlysis of diesel flame by picture processing. JSME International Journal. 32 (3), 434-442.
Shiozaki, T., Suzuki, T., & Shimoda., M. (1980) Observation of combustion process in D.I. diesel engine via high speed direct and schlieren photography. Society of Automotive Engineers. 800025.
Shore, P.R. (1984) The analysis of polycyclic aromatic hydrocarbon emissions. Report by Ricardo Consulting Engineers. Report
Shore, P.R. (1986) Alternative fuel effects on diesel engine PAH and nitro-PAH. Report by Ricardo Consulting Engineers. Report
Shundoh, S., Komori, M., Tsujimura, K, & Kobayashi, S. (1992) NO. reduction from diesel combustion using pilot injection with high pressure fuel injection. Society of Automotive Engineers. 920461.
256
Signer, M. & Steinke, R.E. (1987) Future trends in diesel engine design and their impact on lubricants. Society of Automotive Engineers. 871271.
Sigvardson, K.W. & Birks, J.W. (1984) Detection of nitro-polycyclic aromatic hydrocarbons in liquid chromatography by zinc reduction and peroxyalate chemiluminscence. Journal of Chromatography. 316, 507-518.
Singal, S.K., Pundir, B.P., & Mehta, P.S. (1993) Fuel spray-air motion interaction in DI diesel engines: a review. Society of Automotive Engineers. 930604.
Smith, 0.1. (1981) Fundamentals of soot formation in flames with application to diesel engine particulate emissions. Prog. Energy Combust. Sci. 7, 275-291.
Snook, M.E., Severson, R.F., Higman, H.C., Arrendale, R.F., & Chortyk, Q.T. (1979) Methods for characterization of complex mixtures of polynuclear aromatic hydrocarbons. Jones, P.W. & Leber, P (Editors). Polynuclear Aromatic Hydrocarbons. Ann Arbor Science Publishers
Sorrell, R.K. & Reding, R. (1985) Analysis of polynuclear aromatic hydrocarbons in environmental waters by high-pressure liquid chromatography. Journal of Chromatography. 348, 655-670.
Springer, K.J. ( 1991) Global what? Control possibilities of C02 and other greenhouse gases. Journal of Engineering for Gas Turbines and Power. 113, 440-44 7.
Stark, G., Stauff, J., Miltenburger, H.G., & Stumm-Fischer, I. (1985) Photodecomposition of 1-nitropyrene and other direct-acting mutagens extracted from diesel-exhaust particulates. Mutation Research. 155, 27-33.
Stenburg, U., Alsberg, T., & Westerholm, R. (1983) Emission of carcinogenic components with automobile exhausts. Environmental Health Perspectives. 47, 53-63.
Stober, W. (1992) Health aspects of combustion gases and diesel exhaust. Journal of Aerosol Medicine. 5 (2), 81-102.
Stone, R. (1992) Introduction to Internal Combustion Engines. MacMillan.
Stradling, R.J., Cowley, L.T., Lange, W.W., & Maillard, C. (1993) The influence of fuel properties and test cycle procedures on the exhaust particulate emissions from light-duty diesel vehicles. Institute of Mechanical Engineers Seminar: Fuels for automotive and industrial engines. 61h-71h April.
Tanabe, H., Yamaguchi, H., Yamamoto, T., & Sato, T.G. (1992) Wall effects on SOF formation. Society of Automotive Engineers. 922211
257
Tancell, P.J., Rhead, M.M., Trier, C.J., Bell, M.A., & Fussey, D.E. (1995a) The sources of benzo( a )pyrene in diesel exhaust emissions. The Science of the Total Environment. 162, 179-186.
Tancell, P.J., Rhead, M.M., Trier, C., Pemberton, R., & Braven, J. (1995b) Diesel combustion of an alkylated polcyclic aromatic hydrocarbon. Fuel. awaiting publication.
Tancell (1995). The origin of polycyclic aromtic hydrocarbons in diesel exhaust emissions. Thesis, University of Plymouth, March.
Tejada, S.B., Zweidinger, R.B., & Sigsby, J.E., Jr. (1982) Analysis of nitroaromatics in diesel and gasoline car emissions. Society of Automotive Engineers. 820775.
Tejada, S.B., Zwedinger. R.B., & Sigsby, J.E., Jr. (1986) Fluorescence detection and identification of nitro derivatives of polynuclear aromatic hydrocarbons by on-column catalytic reduction to aromatic amines. Analytical Chemistry. 58, 1827-1834.
Theobald, N. (1988) Rapid preliminary separation of petroleum hydrocarbons by solid-phase extraction cartridges. Analytica Chimica Acta. 204, 135-144.
Tokiwa, H., Nakagawa, R., Kunimasa, M., & Ohnishi, Y. (1981) The mutagenicity of nitro derivatives induced by exposure of non-mutagenic compounds to nitrogen dioxide. Mutation Research. 85, 195-205.
Tokiwa, H. & Ohnishi, Y. (1986) Mutagenicity and carcinogenicity of nitroarenes and their sources in the environment. CRC Critical Reviews in Toxicology. 17 (1), 23-60.
Tomkins, B.A., Brazell, R.S., Roth, M.E., & Ostrum, V. (1984) Isolation of mononitrated polycyclic aromatic hydrocarbons in particulate matter by liquid chromatography and determination by gas chromatography with the thermal energy analyzer. Analytical Chemistry. 56, 781-786.
Tong, H.Y., Sweetman, J.A., & Karasek, F.W. (1983) Quantitation of 1-nitropyrene in diesel exhaust particulates by capillary gas chromatography-mass spectrometry and capillary gas chromatography. Journal of Chromatography. 264, 231-239.
Tong 1984, H.Y., Sweetman, J.A., Karasek, F.W., Jellum, E., & Thorsrud, A.K. (1984) Quantitative analysis of polycyclic aromatic compounds in diesel exhaust extracts by combined chromatographic techniques. Journal of Chromatography. 312, 183-202.
Trier, C.J. (1988) The origins of the organic fraction in diesel exhaust emissions. Thesis, University of Plymouth, June.
258
Trier, C.J., Fussey, D.E., Petch, G.S., & Rhead, M.M. (1988) A comparison of diesel exhaust emissions collected by an environmental protection agency recommended dilution tunnel/filter with the sample from a new solvent scrubbing system. Institute of Mechanical Engineers. C64, 135-144
Trier, C.J., Rhead, M. M., & Fussey, D.E. (1990) Evidence for the pyrosynthesis of parent polcyclic aromatic compounds in the combustion of a diesel engine. Institute of Mechanical Engineers. C394/003, 53-60.
Trier, C.J., Fussey, D.E., Ryder, D., & Graham, M.A. (1991) Developmentofradiotracer techniques for diesel fuel and exhaust emission research. Institute of Mechanical Engineers. C433/010, 159-164.
Trijonis, J. (1984) Effect of diesel vehicles on visibility in California. The Science of the Total Environment. 36, 131-140.
Tritthart, P., Cichiocki, R., & Cartellieri, W. (1993) Fuel effects on emissions in various test cycles in advanced passenger car diesel vehicles. Society of Autmotive Engineers. 932684.
Tsunemoto, H., Yamada, T., & Ishitani, H. (1986) Behaviour of adhering fuel on cold combustion chamber wall in direct injection diesel engines. Society of Automotive Engineers. 861235.
Uden, P.C. (1987) Detection in quantitative gas chromatography. Katz, E. (Editor) Quantitative analysis using chromatographic techniques. John & Wiley Sons, 99-155.
Viegl, E., Posch, W., Linder, W., & Tritthart, P. (1994) Selective and sensitive analysis of 1-nitropyrene in diesel exhauast particulate extract by multidimensional HPLC. Chromatographia. 38 (3/4), 199-206.
Viras, L.G., Athanasiou, K., & Siskos, P. (1990) Determination of mutagenic activity of airborne particulates and of the benzo( a )pyrene concentrations in Athens atmosphere. Atmospheric Environment. 24B (2), 267-274.
Wade, J.F. & Newman, L.S. (1993) Diesel asthma. Reactive airways disease following overexposure to locomotive exhaust. JOM. 35 (2), 149-154.
Waldenmaier, D.A., Gratz, L.D., Bagley, S.T., Johnson, J.H., & Leddy, D.G. (1990) The influence of sampling conditions on the repeatability of diesel particulate vapor phase hydrocarbon and PAH measurements. Society of Automotive Engineers. 900642.
Walsh, M.P. & Bradow, R. (1991) Diesel particulate control around the world. Society of Automotive Engineers. 910130.
Weast, R.C. & Astle, M.J. (1985) CRC Handbook of data on organic compounds. Vol. I & ll, CRC Press.
259
Weaver, C.S., Miller, C., Johnson, W.A., & Higgins, T.S. (1986) Reducing the sulfur and aromatic content of diesel fuel: costs, benefits, and effectiveness for emission control. Society of Automotive Engineers. 860622.
Wei, E.T. & Shu, H.P. (1983) Nitroaromatic carcinogens in diesel soot: a review of laboratory findings. AJPH. 73 (9), 1085-1088.
Weildmann, K., Menrad, H., Reders, K., & Hutcheson R.C. (1988) Diesel fuel quality effects on exhaust emissions. Society of Automotive Engineers. 881649.
Wheeler, R.H. (1984) Diesel exhaust emissions. Haddad, S.D. & Watson, N. (Editors) Principles and performance in diesel engineering. John Wiley & Sons. Chapter 6.
White, C.M., Robbat, A., Jr., & Hoes, R.M. (1984) Evaluation of a thermionic ionization detector for nitrated polycyclic aromatic hydrocarbons. Analytical Chemistry. 56, 232-236.
White, C. M. (1985). Nitrated polycyclic aromatic hydrocarbons. Huethig.
Williams, P.T., Bartle, K.D., & Andrews, G.E. (1986a) The relation between polcyclic aromatic compounds in diesel fuels and exhaust particulates. Fuel. 65, 1150-1158.
Williams, P.T., Andrews, G.E., & Bartle, K.D. (1987) The role of lubricating oil in diesel particulate and particulate PAH emissions. Society of Automotive Engineers. 872084.
Williams, P.T., Abbass, M.K., Andrews, G.E., & Bartle, K.D. (1989) Diesel particulate emissions: the role of unbumed fuel. Combustion and Flame. 75, 1-24.
Williams, P.T. (1990) Sampling and analysis of polycyclic aromatic compounds from combustion systems: a review. Journal of the Institute of Energy. March, 22-30.
Williams, R.L. & Swarin, S.J. (1979) Benzo(a)pyrene emissions from gasoline and diesel automobiles. Society of Automotive Engineers. 790419.
Williams, R.L., Perez, J.M., & Griffing, M.E. (1985) A review of sampling condition effects on polynuclear aromatic hydrocarbons (PNA) from heavy-duty diesel engines. Society of Automotive Engineers. 852081.
Williams, R.L., Sparacino, C., Petersen, B., Bumgame, J., Jungers, R.H., & Lewtas, J. (1986b) Comparative characterization of organic emissions from diesel particles, coke oven mains, roofing tar vapours and cigarette smoke condensate. Intern. J. Environ. Anal. Chem. 26, 27-49.
Williams, R.L., Meares, J., Brooks, L., Watts, R., & Lemieux, P. (1994) Priority pollutant PAH analysis of incinerator emission particles using HPLC and optimized fluorescence detection. Inter. J. Anal. Chem. 54, 299-314.
260
Wise, S.A., Chelser, S.N., Hilpert, L.R., May, W.E., Rebbert, R.E., Vogt, C.R., Nishioka, M.G., Austin, A., & Lewtas, J. (1985) Quantification of polycyclic aromatic hydrocarbons and nitro-substituted polycyclic aromatic hydrocarbons and mutagenicity testing for the character of ambient air particulate matter. J. Environ. lnt. 11, 147-160.
Xu, X.B., Nachtman, J.P., Rappaport, S.M., Wei, E.T., Lewis, S., & Burlingame, A.L. (1981) Identification of 2-nitrofluorene in diesel exhaust particulates. Journal of Applied Toxicology. 1 (3), 196-198.
Xu, X.B., Nachtman, J.P., Jin, Z.L., Wei, E.T., & Rappaport, S.M. (1982) Isolation and identification of mutagenic nitro-PAH in diesel-exhaust particulates. Ana/ytica Chimica Acta. 136, 163-174.
Yergey, J.A., Risby, T.H., & Letz, S.S. (1982) Chemical characterization of organic adsorbates on diesel particulate matter. Analytical Chemistry. 54, 354-357.
Yoshida, E., Nomura, H., & Sekimoto, M. (1986) Fuel and engine effects on diesel exhaust emissions. Society of Automotive Engineers. 860619.
Yu, M. & Rites, R.A. (1981) Identification of organic compounds on diesel engine soot. Analytical Chemistry. 53 (7), 951-954.
Yu, R.C, Wong, V. W, & Shaded, S.M. (1980) Sources of hydrocarbon emissions from direct injection diesel engines. Society of Automotive Engineers. 800048.
Yu, W.C., Fine, D.H., Chiu, K.S., & Biemann, K. (1984) Determination of nitrated polycyclic aromatic hydrocarbons in diesel particulates by gas chromatography with chemiluminescent detection. Analytical Chemistry. 56, 1158-1162.
Zelenka, P., Kriegler, W., Herzog, P.L., & Cartellieri, W.P. (1990) Ways toward the clean heavy-duty diesel. Society of Automotive Engineers. 900602.
Zelenka, P. & Herzog, P.L. (1993) Exhaust gas aftertretment systems for diesel engines with respect to future emission legislation. Institute of Mechanical Engineers Seminar: Worldwide engine emissions and how to meet them. 123-135.
Zhang, L., Minami, T., Takatsuki, T., & Yokota, K. (1993) An analysis of the combustion of a DI diesel engine by photograph processing. Society of Automotive Engineers. 930594.
Zielinska, B., Arey, J., Atkinson, R., & Winer, A.M. (1989a) The nitroarenes of molecular weight 247 in ambient particulate samples collected in southern California. Atmospheric Environment. 23 (I), 223-229.
Zielinska, B., Arey, J., Atkinson, R., & McEiroy, P.A. (1989b) Formation of methylnitronaphthalenes from the gas-phase reactions of 1- and 2-methylnaphthalene with OH radicals and NPs and their occurence in ambient air. Environ. Sci. Technol. 23, 723-729.
261
Ziejewski, M., Goettler, H.J., Cook, L.W., & Flicker, J. (1991) Polycyclic aromatic hydrocarbon emissions from plant oil based alternative fuels. Society of Automotive Engineers. 911765.
Ziejewski, M. & Goettler, H.J. (1992) Comparative analysis of the exhaust emissions for vegetable oil based alternative fuels. Society of Automotive Engineers. 920195.
Zierock, K.-H., Rothe, G., & Steppat, R. (1983) Polycyclic aromatic hydrocarbons in the particulate emissions of three diesel engines. Society of Automotive Engineers. 830458
262
Appendix A - GC-ECD aaalysls of Nllro-PAC coUected usiu& UPCJ1lded TESSA
1-- "'1'"'"""~· (1-~lfD) ....._..., y•t!0.69401;..(),0C.S ~ y-ar-.. &. a-.., kv--t (Ad
...... o ....
- 1'ES (me) --- Wo >ojo<oo4 Mu>. o.n. vo&. (ul) &¥.Vol Arall --.n A•. +0.00 k/60.69140 zDU. Vol. hi: ell ... 1..-tl'l!S A•. TboobiOOml Slla'l'llar Voklmo cll'b.a CAliUIDid MMooiJI>ol MMo ol 1- Av. cl I....., o/Tl!S Wt. ftom. onlfPLC ...... (QI) Ana (b) ....... (nrlmr) "'- olbliDt:. - -- """<-- .... ... ....... .......... .-... [lnc) (IColl Tl!S - ......... (- - o.l6)) - - -<mrl <mrl - (- O<z) 1- {mf/1'1) {mfll'l)
ll<ll ... Muld.
(Ub)& (Ut(~ OL+Ofl 0 a:;Q71) ktor <•) - (rlq) :a:lOO (.u 0 I6J (o/4) -(I) (b) (C) (d) (o) Ill IV ()o) (I) <D 00 (I) (oa) lo) (o) 11>) (q) (Jl (~ (I) ('>) M (W!
1500 1 1.6 11.6 ).$ ... 1.2> 100 0.5 49.1173 49,,.;)6 o49..S9JO 49.6440 0.1119 l lll.>m :IOU 9.5 l6S.41 lO 1).68 O.Oill o.cxoo 0.069J
1500 26.9 26.9 4.7 •.m 1.11 100 0.$ .51.9l00 61..5414 6J.1UJ 63.10 1.0109 11o.r~ US.9 9.1 1!9.66 lO ll.JO o.oco o.cxoo O.OD> 0.07S6
'·' 1500 J3.5 >U H s.c 1.05 100 0.5 u1.1a 121.1916 1)(),,.., IJO.I3ll l.US6 CJI, 11'1'4 4Sl.7 U.$ Ul.U lO n.74 0 . .,.. O.OO'Jl 0.06U
1500 41.4 41.4 11.6 10. 1 l.IS 100 0.5 13.S.65ll 124.2041 1251.9lSIO 129.97«) 2.141.5 4lJ.2924 491.0 10.2 1.14.05 lO n.46 o.am 0.0010 0.01()1 0.06!2
11.1 1500 .... .... 10 .1 9.l 1.10 100 0.5 W .Olll 122..2711 lll. ... l lll.10ll 2.0217 404.131!5 ..... 10.1 96.1 lO ll.11 ...... O.Oill 0.0)99
1500 49.$ 49.$ ll.6 9.1 1.40 100 0.5 1Ullt91 119.1671 119.1Sil 11.9.~1 1.96'73 39l.459l Sl0.8 11.1 96.13 lO Sl.OI O.Q4<t9 0.0112 0.0491 0.044>
10,6
1500 83.2 Sl 1..1 1. 1 1.68 lOO 0.$ lS . .U'JI )1.121.9 36.2799 36.3249 0.$91S .!91.4916 1005.5 ll.l 120 so 41 .67 O.OJ«> 0.0090 O.J i l8
1500 107.2 SO.$ 1.7 4 4.02 100 o.s 511.6"196 104. 1.6&5 IOl.S.Ml 101.43510 1.6713 )34.261.1 Wl.O 14.S 111.S so 4l.SS O,OJISl 0.0092 0.1619 O.IAO< U.J
1500 121.1 34.5 10.1 9.1 2.46 lOO 0..1 <1.!1077 ....... 45.:wa 65.)108 L0761 l076.06f0 2647.1 21.1 10 so 1Ul 0.0616 0.0154 0.1118
1500 Ul.7 ,._, ll.3 uu l.OI lOO 0.5 61.016J .S9.6G2. fO .. U9J 8l.31U 0.9949 994.11965 2686.2 :10.1 (/} so 11.46 0.002> O.OlS6 0.1718 0..1118 :10.9
1500 S7 46.9 IH 10.2 1.1:1 lOO 0.5 57.2051 66.1191 62.012S 62.Q51S 1.022> 1112lA6U 1111.1 .12.1 M .U so 1:43 .• 0.11Al O.OllO O.OOOl
1500 .... S0.6 11.6 ... 2.43 lOO 0.5 Q...1111 65..1612 &l.lo!C 6 .9112 l .QSJO 1053.000 :!369.! JU J4.61.5 so I .US O. llA7 0.0!12 0.001 0.0714 I>. I
A 1
I Appendix A- GC-ECD analy&is of Nit.r.,..PAC coUected usin& Upgraded TESSA
J-m<tkyl-1-rfilrorutpllllwh•• (2-M•·I·•M) - y-69.4&51x+0.269 wt:w:ny-ara&x-~M~(n&)
R«l•0.991 - TES (mol -- - Wt m;ec.s ..... Dil. Vol. (\1.1) q . Vul Anal! """" Av. j<>.269 k169.41J7 J:Dil. Vol. _,._o(u..t.t ... ,._,....m!$ Av. TU..btOOml - Volume o( lW ec...am.d """'"' ... ...... ol U.-1~ Av. cCl-M.-1...-
o/TES W1.fraal cnHPLC ..... ..n Ano(b) ....... .,._ ol- ala.!. ID b. - -- --- ""' .... _. .... - .... .,....., - ()hDTES - - (..,) - 0.1163) - - -(awl - - (-.) 011) ·- Cmlll<ll <moll<ll
Qlz) ......... (c:Jb) X (8)K (1/d) Qo+M 0 x(fld faclol" (e) ""'" (r)Q)x 100 (sxO.III.J) (uol) -(o) (b) (c) (d) (o) 00 (<) 01) (I) (I) 00 (I) (IQ) (o) (o) (p) (q) (I! (o) (1) (11) (., ("')
l>()() 21.6 21.6 >.5 4 .. 1.2> 100 0..1 46.%nl 49.JJU ....... .. ...,. o ..... 136-'014 110.6 1.9 3&5.41 "' u ... 0.0111 0.00!0 O.OS71 uoo ,.., 26.9 4.7 4.03 1.17 100 0.> ... .,., 11.1118 61.$161 61.!011 0.9017 1.91.7165 2267 1.4 J.S9.66 "' U.90 0.0120 0.00.!0 0.07.S6
1.2 0.06fl1
uoo >>.> ll..l >.9 ,.., 1.05 100 O.> 108..1124 99.6101 10U)6JA 103.792A 1.4911 291.1.131 JlJ.7 9.4 l.ll.1l "' 12.74 O.OZil 0.0011 0.014< uoo 41.4 41.4 11.6 10.1 I.U 100 0..1 111.9577 1.12.09.12 10>.02» 161.7.565 l.lliO <16.5.S9:Sol ll$.9 11.1 U4.05 "' 32.46 O.OZIO 0.0010 0.0765
10.2 0.06QS
uoo .... 44.1 10.1 9.2 1.10 100 0..1 101.6166 11!2.116C2 103.1A04 102..9714 1.4120 296.3Sl01 >26.0 7.4 96.1 "' .51.71 0.0446 0,0112 o.am l>()() 49 . .5 49 . .5 u.o 9.1 .... 100 0.> 109..100 ...... ... = 98.911.5 1.4231 214.76U 391.7 1.1 96.13 "' .52.01 0,0449 0.0112 o.om
1.1 0.0.12A
»00 tl.l >2 1..1 8.1 .... >00 0.> l4.1QU 22.7 .. 'lJ.42Al 2>. W7 O.ll3J 333.2ll0 1S9.9 '·' lllO "' 4l.6l 0.0>00 O.o090 o ..... »00 107.2 lO..I 1.7 4 4.112 100 o.> ... ..,. 14.1141. 11.Dl 11.1l9J 1.0296 205JI161 liS> ..I 1.9 111.$ "' 42..1> 0.0167 0.0092 O.lOU
1.1 o.om »00 121.7 >4..1 10.1 9.1 2.46 >00 0..1 J7.2923 u.nn ><.OUO 17.7440 0..14>2 .... ~ 11l6.J 11.0 70 "' 71.43 0.0616 O.Ol.Sf 0.0167
»00 U1.l >9..1 12.1 lO.J 2.68 >00 >00 44.20.12 42.1011 C .liS8 42.9161 0.6177 617.mt 1667.1 12..1 .. "' 12.46 0.0112> O.Ol.S6 0.1061
11.1 0.0967 »00 >1 46.9 U.l 10.2 1.112 >00 0..1 116.9'114 lll.l467 124.109.1 llJ.ItOI 1.7CU :1>6.4>11 11>2.1 11.4 14.7$ "' 141 .• 0.1242 0.0110 o.cnto »00 .... "'·' 11.6 9.1 2.41 >00 0..1 2.46.0106 2>4.041 2>0.029'.1 249.1601 1..1045 1 11.90Jl 1154.1 2>.6 3.4.615 "' 144.4.5 0.1247 0.0112 O.ll>fJ
11..1 0.0386
A2
I I Appendix A- GC-ECD analysis of Nilro-PAC coUe<ted using Upgp&ded TESSA
I j j R«t•0.997
M ... OiL Val (1ll) lnJ, Vol +0.158 1 DD. Vol. .,ot:z.... Av. oll<ID/
ol TES Wt. &am m...llPLC ...... (ulj ........ , _ ... , -- "'""- """' (jfallTP.S
- (m:) '"''"""' mxMulli.
(elb) x (-.)x (lld) a x (JirJ &ca (e) (ri<IJ xlOO (&1: 0.863) V() -(b) (e) (dJ (e) 00 (i) (j) (I) (m) (o) (p) (q) (rj b) (I)
2 1.6 21.6 4.4 l .lS 100 O.J 29.0739 31..4471 30.2605 30.4L8S 118.1319 147.7 6.8 3155.41 1~.68 0.0118 0.0010 O.OSOO
26.9 16.9 4.7 • . Ol 1.11 100 19.6199 46SW 43.1012 43.2592 0.8400 161.9995 196.6 7.1 159.66 13.90 0.0120 O.OOJO 7.1 0.0578
5.9 5.62 !.OS 100 O.J 59.2603 57 . .5Ui9 51.SIB6 58.,So466 1.1368 '227.3690 218.7 7.1 1.51.71 o.ow 0.0071 o.om 48.4 48.4 11.6 10.1 I.U 100 O.J 57.5697 .5l.j2l 54.5454 .S'-10).4 1.0622 5. 1 ll< .OS 12.46 0.02lkl 0.0070 0.0)49
6.1 0.0!44 .... 10. 1 9.l 1. 10 100 a>.7454 79.6844 1).2149 liU729 I.J607 3-41.1 7.8 96.7 51.71 0.0446 0.0112 0.0308
49.> 49.> IJ.6 9.7 1.40 100 0.5 71.1391 59. 1861 65.5126 65.6706 1.2751 ID.O 7.l 96.Jl 52.01 0.0449 0 .0112 0,0318
0.0'111
8l.2 ll 8.5 S.J .... 14.)999 l6.9U8 20.612:9 20.009 0.40<5 619.5 O.l 110 41.61 0.03(1(1 O.OOSIO 0.0156 107.2 8.7 .. ., 100 ))..5602 1S. I914 34-.31'!H l4..S3'7S M!Otl 022.4 5.1 117.5 0.0167 0.009l 0,()578
7.0 0 .(1717
121.1 10.7 9.7 l ... O.J 22.52.51 2J . .c2J9 '22.9745 23.1325 0.4492 449. 1811 1105.0 9. 1 70 11.45 0.0616 0 .0154 0.07l7
113.7 ll.l 10.3 2.61 16.9929 16.8949 16.9439 17.1019 0.3321 896.6 6.1 .. 72.46 0.0625 0.0156 O.OS74 1.9
l>OO >1 46.9 U .l 10.2 1.11! 100 O.J 21.6791 2A .. 92S6 23.30'27 23.4607 0.4l56 8.H.7 14.6 34.75 143.88 0. 12A2 0.0310 0 .0269
61.6 17.6 9.8 O.J 31.8602 )4.0333 33.4461 33.6048 0.6525 IJ9l.l 2l.2 14.615 144.45 0.12A1 0.().112 O.Q511
18.9 0.0390
A3
I Appeocli A- GC-ECD aaaJ;rsls oC Nltre>-PAC collected usiog Upexaded TESSA
1-lfitropyro•• (I -• ~ ~ y•l&.OJ69a.+O. M ...t.:n y--. A :r.- wc. illjec:lod Otc)
R«!-O.Il9!
- TES Cmrl -- - w,_ Mull 011 Vol. (bO lnj . Vol Atall .._n Av. 14.6! kf18.0.s69 :r.DD. Vol. .. ae,..., 1....rt1!S Av. Tbnab lOOinl - v-.n. ot. Plld eao.um.s. Ma.. oi..!Jel Muo oC 1- Av, a/l-ap/
oC TBS WL- oa H:PLC """" (Ill) Ala (b) ....... 0../md o(- albiiOba - -- uood (- - .... ""' ........ ......... ....... (m<) (lhDTES lppm) - <- > - 0 .1163) - - ........ (me) - .- (- Qar) 1 <Y- (molb) ""'""'' 0<21
m •Muld. (c:JtUa W• (Jld) Ol+ ;jll 0. (fir) 6dor <•> - (rfq):r.IOO •x0.163 (114) OW)
c.> (b) (c) (d) (e) (l) (I) Ol) (I) U> 00 (l) (1nl (1>1 (o) (p) (a)
~· w (Q (11) (v) (IN)
1.500 21.6 21. 6 ,, ... 1.1! 100 0.> 6.t152 UWOl 6.!Am >.&m 0-""0 l0.7991J "'" ... lM.oil! ,. U .61 0.0111 O.OOlO O.OllO
1.500 U.P 26.9 '·' .... 1. 11 100 O.> s.mt 7.9)34 .. ,., 6. 1l0l 0.16J.l = n.• 1.4 lS9.66 ,. U .90 0.0120 O.OOlO o.ow 1.6 0.0120
1.500 11.> JU 5.9 ,., 1.0> 100 O,j ....... ··- 6.6461 1.9961 0. 1516 n.ms .51.1 1.0 U2.7J ,. Jl.74 0.0210 0.0071 o.oon L500 44 44 11.6 10. 1 I. U 100 0.> 9.1511 .. u.c.w: 10.>111 ...... 0.1!11.1 11.1821 59.6 1.2 U4.0l ,. Jl.44 0.0200 o.rmo o.ooos
1.1 0.0066
1.500 .. ,, 44. 1 10. 1 0.2 I.JO 100 o.> lC.SVZ:Z 15. UG. l.tll..&in IA.:.>!2 O.l1:U 74.6917 12.0 L9 01!.7 ,. .51~11 0.0446 0.0112 0.0074
1.500 .. , .. , U .6 9.1 I .AO 100 O.> 3.6615 >.12611 5.C9Q 4.144 0~1.%7" 25.<1701 3>.7 0.7 OIS.U ,. Jl.OI 0.0449 o.ouz 0.0032
I.J O.OO!l
2>00 >9.1 ,, lJ 10 ,,.. 140 0.> ,,... 3.0006
, __ l...., 0.0601S lJ.•m ., .. L1 1.9>.24 "' 2Ul 0.022.1 0 .00» 0.0122
2>00 ,, JI.J "'·' I 4,64 100 0.> l.!C!1 l.f79S l.- 2.1!01 O.GSI6 J1 .72SN 54.4 0.9 IJ7l.P "' 2>.l7l 0.022A 0 ..... 0.0001
1.0 0.0109 2>00 7 L7 3>.9 2Uil 10 4.J1 100 0.> ...... ... .,. ....... 7.75111 O.lOOO •U.OO"J:: 17SU l.J IJ "' 6L!l 0.051J 0 .013) o.ow 2>00 9').6 44.1 2>. 1 10 >.02 tOO 0.> 9.1091 10.6C1S 9.1'113 ·-- 0.2426 41.>229 2A.U 2.6 ., ,. ., .. O.QS215 O.OU 1 0.0115
2.6 0.0160
2>00 12.1 21.2 11.1 10 5.21 100 O.> 1.»02 O.U02 um 7.67S2 O.lOll 40.1>66 212.7 l.J Sl.l "' ..... o.om O.O'XIL O.OIOd
2>00 91.2 U .4 19 .1 0.9 J.ll 100 0 .> 7.0619 .S.W7 ...... ,.. ... 0.1411 li.611S 166.2 l.O .. ,, "' 91.96 0.177 .. O.Dl!ill 0.0014
l.2 0 .009.1
J>OO 1.1.2 >2 1.> 1.1 1.61 100 0 .> 7.7L9l 1 .0043 7.l 618 6.1111 0 . 1765 l$.1910 59.l 0.1 120 "' 41.67 O.lllOO 0.0090 0.0066
J>OO 107.'2 "'" 1.7 • 4.62 100 0.> '2.7lll l.PI2 2.1121 2.1QB 0.0569 ll.lll9 = O.> 11'7.J "' ....., 0.0161 0 .0002 O.OQJ7
0 .6 0 .0062
J>OO 121.7 .. , 10,1 0.1 l.46 lOO 0 .> 6.17721 6.UI.I 6.2116 $.>616 0. 1464 29.2000 n.l 0.6 10 "' 71.43 0 .0616 O.OLSC 0.0047
J>OO ll5.7 ,., 12.3 lO.l l.61 >00 0.> 2.5989 1.1201 1.9595 I. :lOP> 0.0 )44 :M.427l Pl.4 0.7 .. "' 72.46 0.0625 0.0156 0.0059
0,6 o.oon J>OO , 46.9 U .J 10.2 1.11 100 0.$ S4.669l )4.7449 14.7071 l4.0S71 0 ..... 179.0140 !.26.> $,1 ""·" "' .... ... 0 .12Cl O.OllO 0.0105
J>OO 61.6 >0.6 11.6 9.8 l.<J 100 O.> 26.0717 U.:l4ll6 26.2137 l.S.S6l7 0.672.1 u.uuo 327.> 4 .8 J4 .6LS ,. ..... ., O. llA7 O.Olll 0.0105
>.2 0.0105
• 4
